diff --git a/darwin-x86/bin/cxx_extractor b/darwin-x86/bin/cxx_extractor
index 450b4f9..62b57c3 100755
--- a/darwin-x86/bin/cxx_extractor
+++ b/darwin-x86/bin/cxx_extractor
Binary files differ
diff --git a/darwin-x86/bin/header-abi-diff b/darwin-x86/bin/header-abi-diff
index c1ecacb..65698c8 100755
--- a/darwin-x86/bin/header-abi-diff
+++ b/darwin-x86/bin/header-abi-diff
Binary files differ
diff --git a/darwin-x86/bin/header-abi-dumper b/darwin-x86/bin/header-abi-dumper
index a3896c7..2fea1d1 100755
--- a/darwin-x86/bin/header-abi-dumper
+++ b/darwin-x86/bin/header-abi-dumper
Binary files differ
diff --git a/darwin-x86/bin/header-abi-linker b/darwin-x86/bin/header-abi-linker
index 482d80b..187aee3 100755
--- a/darwin-x86/bin/header-abi-linker
+++ b/darwin-x86/bin/header-abi-linker
Binary files differ
diff --git a/darwin-x86/bin/merge-abi-diff b/darwin-x86/bin/merge-abi-diff
index db7c4b0..429357f 100755
--- a/darwin-x86/bin/merge-abi-diff
+++ b/darwin-x86/bin/merge-abi-diff
Binary files differ
diff --git a/darwin-x86/bin/proto_metadata_plugin b/darwin-x86/bin/proto_metadata_plugin
new file mode 100755
index 0000000..53175a5
--- /dev/null
+++ b/darwin-x86/bin/proto_metadata_plugin
Binary files differ
diff --git a/darwin-x86/bin/protoc_extractor b/darwin-x86/bin/protoc_extractor
index 895a982..87af276 100755
--- a/darwin-x86/bin/protoc_extractor
+++ b/darwin-x86/bin/protoc_extractor
Binary files differ
diff --git a/darwin-x86/bin/versioner b/darwin-x86/bin/versioner
index e117527..73d949b 100755
--- a/darwin-x86/bin/versioner
+++ b/darwin-x86/bin/versioner
Binary files differ
diff --git a/darwin-x86/clang-headers b/darwin-x86/clang-headers
index a5cbaaf..4009f74 120000
--- a/darwin-x86/clang-headers
+++ b/darwin-x86/clang-headers
@@ -1 +1 @@
-lib64/clang/9.0.7/include
\ No newline at end of file
+lib64/clang/10.0.1/include
\ No newline at end of file
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_builtin_vars.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_builtin_vars.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_builtin_vars.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_builtin_vars.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_cmath.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_cmath.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_cmath.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_cmath.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_complex_builtins.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_complex_builtins.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_complex_builtins.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_complex_builtins.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_device_functions.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_device_functions.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_device_functions.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_device_functions.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
similarity index 98%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
index 2970d17..b67461a 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
@@ -211,7 +211,15 @@
   return __nvvm_vote_ballot_sync(mask, pred);
 }
 
-inline __device__ unsigned int __activemask() { return __nvvm_vote_ballot(1); }
+inline __device__ unsigned int __activemask() {
+#if CUDA_VERSION < 9020
+  return __nvvm_vote_ballot(1);
+#else
+  unsigned int mask;
+  asm volatile("activemask.b32 %0;" : "=r"(mask));
+  return mask;
+#endif
+}
 
 inline __device__ unsigned int __fns(unsigned mask, unsigned base, int offset) {
   return __nvvm_fns(mask, base, offset);
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_libdevice_declares.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_libdevice_declares.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_libdevice_declares.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_libdevice_declares.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_math_forward_declares.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_math_forward_declares.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_math_forward_declares.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_math_forward_declares.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_runtime_wrapper.h b/darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_runtime_wrapper.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_runtime_wrapper.h
rename to darwin-x86/lib64/clang/10.0.1/include/__clang_cuda_runtime_wrapper.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__stddef_max_align_t.h b/darwin-x86/lib64/clang/10.0.1/include/__stddef_max_align_t.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__stddef_max_align_t.h
rename to darwin-x86/lib64/clang/10.0.1/include/__stddef_max_align_t.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__wmmintrin_aes.h b/darwin-x86/lib64/clang/10.0.1/include/__wmmintrin_aes.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__wmmintrin_aes.h
rename to darwin-x86/lib64/clang/10.0.1/include/__wmmintrin_aes.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__wmmintrin_pclmul.h b/darwin-x86/lib64/clang/10.0.1/include/__wmmintrin_pclmul.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/__wmmintrin_pclmul.h
rename to darwin-x86/lib64/clang/10.0.1/include/__wmmintrin_pclmul.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/adxintrin.h b/darwin-x86/lib64/clang/10.0.1/include/adxintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/adxintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/adxintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/altivec.h b/darwin-x86/lib64/clang/10.0.1/include/altivec.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/altivec.h
rename to darwin-x86/lib64/clang/10.0.1/include/altivec.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ammintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ammintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/ammintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ammintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm64intr.h b/darwin-x86/lib64/clang/10.0.1/include/arm64intr.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/arm64intr.h
rename to darwin-x86/lib64/clang/10.0.1/include/arm64intr.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm_acle.h b/darwin-x86/lib64/clang/10.0.1/include/arm_acle.h
similarity index 94%
rename from darwin-x86/lib64/clang/9.0.7/include/arm_acle.h
rename to darwin-x86/lib64/clang/10.0.1/include/arm_acle.h
index 08d65fa..0510e6f 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/arm_acle.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/arm_acle.h
@@ -597,6 +597,14 @@
 }
 #endif
 
+/* Armv8.3-A Javascript conversion intrinsic */
+#if __ARM_64BIT_STATE && defined(__ARM_FEATURE_JCVT)
+static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
+__jcvt(double __a) {
+  return __builtin_arm_jcvt(__a);
+}
+#endif
+
 /* 10.1 Special register intrinsics */
 #define __arm_rsr(sysreg) __builtin_arm_rsr(sysreg)
 #define __arm_rsr64(sysreg) __builtin_arm_rsr64(sysreg)
@@ -605,7 +613,7 @@
 #define __arm_wsr64(sysreg, v) __builtin_arm_wsr64(sysreg, v)
 #define __arm_wsrp(sysreg, v) __builtin_arm_wsrp(sysreg, v)
 
-// Memory Tagging Extensions (MTE) Intrinsics
+/* Memory Tagging Extensions (MTE) Intrinsics */
 #if __ARM_FEATURE_MEMORY_TAGGING
 #define __arm_mte_create_random_tag(__ptr, __mask)  __builtin_arm_irg(__ptr, __mask)
 #define __arm_mte_increment_tag(__ptr, __tag_offset)  __builtin_arm_addg(__ptr, __tag_offset)
@@ -615,6 +623,28 @@
 #define __arm_mte_ptrdiff(__ptra, __ptrb) __builtin_arm_subp(__ptra, __ptrb)
 #endif
 
+/* Transactional Memory Extension (TME) Intrinsics */
+#if __ARM_FEATURE_TME
+
+#define _TMFAILURE_REASON  0x00007fffu
+#define _TMFAILURE_RTRY    0x00008000u
+#define _TMFAILURE_CNCL    0x00010000u
+#define _TMFAILURE_MEM     0x00020000u
+#define _TMFAILURE_IMP     0x00040000u
+#define _TMFAILURE_ERR     0x00080000u
+#define _TMFAILURE_SIZE    0x00100000u
+#define _TMFAILURE_NEST    0x00200000u
+#define _TMFAILURE_DBG     0x00400000u
+#define _TMFAILURE_INT     0x00800000u
+#define _TMFAILURE_TRIVIAL 0x01000000u
+
+#define __tstart()        __builtin_arm_tstart()
+#define __tcommit()       __builtin_arm_tcommit()
+#define __tcancel(__arg)  __builtin_arm_tcancel(__arg)
+#define __ttest()         __builtin_arm_ttest()
+
+#endif /* __ARM_FEATURE_TME */
+
 #if defined(__cplusplus)
 }
 #endif
diff --git a/darwin-x86/lib64/clang/10.0.1/include/arm_fp16.h b/darwin-x86/lib64/clang/10.0.1/include/arm_fp16.h
new file mode 100644
index 0000000..1c04f2a
--- /dev/null
+++ b/darwin-x86/lib64/clang/10.0.1/include/arm_fp16.h
@@ -0,0 +1,590 @@
+/*===---- arm_fp16.h - ARM FP16 intrinsics ---------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ARM_FP16_H
+#define __ARM_FP16_H
+
+#include <stdint.h>
+
+typedef __fp16 float16_t;
+#define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
+
+#if defined(__ARM_FEATURE_FP16_SCALAR_ARITHMETIC) && defined(__aarch64__)
+#define vabdh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vabsh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
+  __ret; \
+})
+#define vaddh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcageh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcagth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcaleh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcalth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vceqh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vceqzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
+  __ret; \
+})
+#define vcgeh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcgezh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
+  __ret; \
+})
+#define vcgth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcgtzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
+  __ret; \
+})
+#define vcleh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vclezh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
+  __ret; \
+})
+#define vclth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcltzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
+  __ret; \
+})
+#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvth_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvth_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
+  __ret; \
+})
+__ai float16_t vcvth_f16_u32(uint32_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_u64(uint64_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_u16(uint16_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s32(int32_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s64(int64_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s16(int16_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
+  return __ret;
+}
+#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
+  __ret; \
+})
+#define vcvtmh_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
+  __ret; \
+})
+#define vdivh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __s2 = __p2; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
+  __ret; \
+})
+#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __s2 = __p2; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
+  __ret; \
+})
+#define vmaxh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vminh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vminnmh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmulh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmulxh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vnegh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
+  __ret; \
+})
+#define vrecpeh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
+  __ret; \
+})
+#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vrecpxh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
+  __ret; \
+})
+#define vrndh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
+  __ret; \
+})
+#define vrndah_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
+  __ret; \
+})
+#define vrndih_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
+  __ret; \
+})
+#define vrndmh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
+  __ret; \
+})
+#define vrndnh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
+  __ret; \
+})
+#define vrndph_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
+  __ret; \
+})
+#define vrndxh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
+  __ret; \
+})
+#define vrsqrteh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
+  __ret; \
+})
+#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vsqrth_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
+  __ret; \
+})
+#define vsubh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
+  __ret; \
+})
+#endif
+
+#undef __ai
+
+#endif /* __ARM_FP16_H */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm_neon.h b/darwin-x86/lib64/clang/10.0.1/include/arm_neon.h
similarity index 86%
rename from darwin-x86/lib64/clang/9.0.7/include/arm_neon.h
rename to darwin-x86/lib64/clang/10.0.1/include/arm_neon.h
index 694bdfc..6b55c4b 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/arm_neon.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/arm_neon.h
@@ -1042,20 +1042,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1124,20 +1115,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -1457,20 +1439,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 & __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1522,20 +1495,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 & __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -1723,20 +1687,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 & ~__p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1788,20 +1743,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 & ~__p1;
   return __ret;
 }
-#else
-__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -2089,20 +2035,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
   uint16x4_t __ret;
@@ -2175,20 +2112,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
   int16x4_t __ret;
@@ -4191,174 +4119,66 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vcreate_p8(uint64_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vcreate_p8(uint64_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vcreate_p16(uint64_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vcreate_p16(uint64_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vcreate_u8(uint64_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vcreate_u8(uint64_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcreate_u32(uint64_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vcreate_u32(uint64_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcreate_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vcreate_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcreate_u16(uint64_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vcreate_u16(uint64_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vcreate_s8(uint64_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vcreate_s8(uint64_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcreate_f32(uint64_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vcreate_f32(uint64_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vcreate_f16(uint64_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vcreate_f16(uint64_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vcreate_s32(uint64_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vcreate_s32(uint64_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcreate_s64(uint64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vcreate_s64(uint64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vcreate_s16(uint64_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vcreate_s16(uint64_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
@@ -4899,22 +4719,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -4987,22 +4797,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -5263,20 +5063,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vdup_n_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vdup_n_u16(uint16_t __p0) {
   uint16x4_t __ret;
@@ -5354,20 +5145,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vdup_n_s64(int64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai int64x1_t vdup_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vdup_n_s16(int16_t __p0) {
   int16x4_t __ret;
@@ -5553,20 +5335,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 ^ __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -5618,20 +5391,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 ^ __p1;
   return __ret;
 }
-#else
-__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -5964,7 +5728,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -5972,16 +5735,6 @@
   __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vext_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -6066,7 +5819,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -6074,16 +5826,6 @@
   __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vext_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -6685,28 +6427,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vget_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -6799,28 +6525,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vget_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -7667,20 +7377,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64(__p0) __extension__ ({ \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16(__p0) __extension__ ({ \
   uint16x4_t __ret; \
@@ -7741,20 +7442,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64(__p0) __extension__ ({ \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16(__p0) __extension__ ({ \
   int16x4_t __ret; \
@@ -7995,20 +7687,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_u64(__p0) __extension__ ({ \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
   __ret; \
 })
-#else
-#define vld1_dup_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_dup_u16(__p0) __extension__ ({ \
   uint16x4_t __ret; \
@@ -8069,20 +7752,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_s64(__p0) __extension__ ({ \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
   __ret; \
 })
-#else
-#define vld1_dup_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_dup_s16(__p0) __extension__ ({ \
   int16x4_t __ret; \
@@ -8368,22 +8042,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -8456,22 +8120,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -8745,20 +8399,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x2(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x2(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x2(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -8827,20 +8472,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x2(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x2(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x2(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -9128,20 +8764,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x3(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x3(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x3(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -9214,20 +8841,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x3(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x3(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x3(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -9531,20 +9149,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x4(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x4(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x4(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -9621,20 +9230,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x4(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x4(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x4(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -9875,20 +9475,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_u64(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld2_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_u16(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -9957,20 +9548,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_s64(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld2_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_s16(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -10243,20 +9825,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_u64(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld2_dup_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_dup_u16(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -10325,20 +9898,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_s64(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld2_dup_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_dup_s16(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -10920,20 +10484,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_u64(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld3_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_u16(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -11006,20 +10561,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_s64(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld3_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_s16(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -11308,20 +10854,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_u64(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld3_dup_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_dup_u16(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -11394,20 +10931,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_s64(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld3_dup_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_dup_s16(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -12033,20 +11561,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_u64(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld4_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_u16(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -12123,20 +11642,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_s64(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld4_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_s16(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -12441,20 +11951,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_u64(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld4_dup_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_dup_u16(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -12531,20 +12032,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_s64(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld4_dup_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_dup_s16(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -14996,20 +14488,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vmov_n_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vmov_n_u16(uint16_t __p0) {
   uint16x4_t __ret;
@@ -15087,20 +14570,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vmov_n_s64(int64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai int64x1_t vmov_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vmov_n_s16(int16_t __p0) {
   int16x4_t __ret;
@@ -16854,20 +16328,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 | ~__p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -16919,20 +16384,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 | ~__p1;
   return __ret;
 }
-#else
-__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -17120,20 +16576,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 | __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -17185,20 +16632,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 | __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -18239,20 +17677,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -18309,20 +17738,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -19676,20 +19096,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -19741,20 +19152,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -20158,20 +19560,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -20223,20 +19616,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -20434,22 +19818,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshl_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshl_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -20504,22 +19878,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshl_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshl_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -20646,22 +20010,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -21076,20 +20430,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -21146,20 +20491,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -22362,20 +21698,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -22427,20 +21754,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -22638,22 +21956,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrshr_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrshr_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -22708,22 +22016,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrshr_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrshr_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -23194,7 +22492,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -23202,16 +22499,6 @@
   __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -23275,7 +22562,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -23283,16 +22569,6 @@
   __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -23851,7 +23127,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -23859,23 +23134,6 @@
   __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16_t __s0 = __p0; \
@@ -23984,7 +23242,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -23992,23 +23249,6 @@
   __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16_t __s0 = __p0; \
@@ -24206,20 +23446,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -24271,20 +23502,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -24482,22 +23704,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshl_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshl_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -24552,22 +23764,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshl_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshl_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -24910,22 +24112,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshr_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshr_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -24980,22 +24172,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshr_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshr_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -25452,7 +24634,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -25460,16 +24641,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -25533,7 +24704,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -25541,16 +24711,6 @@
   __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -25782,7 +24942,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -25790,16 +24949,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -25863,7 +25012,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -25871,16 +25019,6 @@
   __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -26196,7 +25334,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -26204,16 +25341,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -26277,7 +25404,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -26285,16 +25411,6 @@
   __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -26511,18 +25627,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
 })
-#else
-#define vst1_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -26575,18 +25683,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
 })
-#else
-#define vst1_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -26795,18 +25895,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
 })
-#else
-#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -26859,18 +25951,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
 })
-#else
-#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -27109,18 +26193,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x2(__p0, __p1) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
 })
-#else
-#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x2(__p0, __p1) __extension__ ({ \
   uint16x4x2_t __s1 = __p1; \
@@ -27181,18 +26257,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x2(__p0, __p1) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
 })
-#else
-#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x2(__p0, __p1) __extension__ ({ \
   int16x4x2_t __s1 = __p1; \
@@ -27448,18 +26516,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x3(__p0, __p1) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
 })
-#else
-#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x3(__p0, __p1) __extension__ ({ \
   uint16x4x3_t __s1 = __p1; \
@@ -27524,18 +26584,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x3(__p0, __p1) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
 })
-#else
-#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x3(__p0, __p1) __extension__ ({ \
   int16x4x3_t __s1 = __p1; \
@@ -27807,18 +26859,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x4(__p0, __p1) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
 })
-#else
-#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x4(__p0, __p1) __extension__ ({ \
   uint16x4x4_t __s1 = __p1; \
@@ -27887,18 +26931,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x4(__p0, __p1) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
 })
-#else
-#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x4(__p0, __p1) __extension__ ({ \
   int16x4x4_t __s1 = __p1; \
@@ -28111,18 +27147,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_u64(__p0, __p1) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
 })
-#else
-#define vst2_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2_u16(__p0, __p1) __extension__ ({ \
   uint16x4x2_t __s1 = __p1; \
@@ -28183,18 +27211,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_s64(__p0, __p1) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
 })
-#else
-#define vst2_s64(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2_s16(__p0, __p1) __extension__ ({ \
   int16x4x2_t __s1 = __p1; \
@@ -28643,18 +27663,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_u64(__p0, __p1) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
 })
-#else
-#define vst3_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3_u16(__p0, __p1) __extension__ ({ \
   uint16x4x3_t __s1 = __p1; \
@@ -28719,18 +27731,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_s64(__p0, __p1) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
 })
-#else
-#define vst3_s64(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3_s16(__p0, __p1) __extension__ ({ \
   int16x4x3_t __s1 = __p1; \
@@ -29208,18 +28212,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_u64(__p0, __p1) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
 })
-#else
-#define vst4_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4_u16(__p0, __p1) __extension__ ({ \
   uint16x4x4_t __s1 = __p1; \
@@ -29288,18 +28284,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_s64(__p0, __p1) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
 })
-#else
-#define vst4_s64(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4_s16(__p0, __p1) __extension__ ({ \
   int16x4x4_t __s1 = __p1; \
@@ -29759,20 +28747,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -29841,20 +28820,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -32085,3702 +31055,1326 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #endif
 #if (__ARM_FP & 2)
 #ifdef __LITTLE_ENDIAN__
@@ -37072,20 +33666,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vsha1h_u32(uint32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vsha1h_u32(uint32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
   uint32x4_t __ret;
@@ -37388,20 +33973,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrndns_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrndns_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vrndpq_f32(float32x4_t __p0) {
   float32x4_t __ret;
@@ -37818,20 +34394,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37848,20 +34415,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37878,20 +34436,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37908,20 +34457,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37938,20 +34478,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37968,20 +34499,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37998,20 +34520,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -38028,5508 +34541,1971 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #endif
 #if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
 #ifdef __LITTLE_ENDIAN__
@@ -43548,20 +36524,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrnd_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrnd_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndaq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43578,20 +36545,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrnda_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrnda_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndiq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43608,20 +36566,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndi_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndi_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndmq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43638,20 +36587,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndm_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndm_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndnq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43668,20 +36608,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndn_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndn_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndpq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43698,20 +36629,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndp_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndp_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndxq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43728,20 +36650,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndx_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndx_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #endif
 #if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
 #ifdef __LITTLE_ENDIAN__
@@ -43761,20 +36674,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -43792,20 +36696,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #endif
 #if defined(__ARM_FEATURE_DOTPROD)
 #ifdef __LITTLE_ENDIAN__
@@ -48073,48 +40968,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vabsq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -48147,48 +41015,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vabs_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vabs_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vabs_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vabsd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -48206,48 +41047,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -48791,20 +41605,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
   return __ret;
 }
-#else
-__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
   poly64x2_t __ret;
@@ -48841,20 +41646,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48872,48 +41668,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48931,48 +41700,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48990,48 +41732,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -49049,62 +41764,26 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
   uint64x2_t __ret;
@@ -49173,104 +41852,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
@@ -49287,20 +41903,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
@@ -49557,20 +42164,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
   uint16x4_t __ret;
@@ -49603,20 +42201,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vceqz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -49649,20 +42238,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -49679,62 +42259,26 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqzd_u64(uint64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vceqzd_u64(uint64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vceqzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vceqzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vceqzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vceqzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vceqzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -49786,104 +42330,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -49996,20 +42477,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgez_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcgez_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50042,20 +42514,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgez_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcgez_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50072,48 +42535,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgezd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcgezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgezd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcgezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgezs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcgezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50165,104 +42601,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -50375,20 +42748,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50421,20 +42785,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50451,48 +42806,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgtzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcgtzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcgtzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgtzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcgtzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50544,104 +42872,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vclezq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -50754,20 +43019,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclez_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vclez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vclez_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50800,20 +43056,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclez_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vclez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vclez_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50830,48 +43077,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vclezd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vclezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vclezd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vclezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vclezs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vclezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50923,104 +43143,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -51133,20 +43290,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcltz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcltz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -51179,20 +43327,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcltz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcltz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -51209,48 +43348,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcltzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcltzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcltzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcltzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcltzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
   poly64x2_t __ret;
@@ -51379,7 +43491,7 @@
   uint64x1_t __s2_21 = __p2_21; \
   uint64x2_t __rev0_21;  __rev0_21 = __builtin_shufflevector(__s0_21, __s0_21, 1, 0); \
   uint64x2_t __ret_21; \
-  __ret_21 = __noswap_vsetq_lane_u64(__noswap_vget_lane_u64(__s2_21, __p3_21), __rev0_21, __p1_21); \
+  __ret_21 = __noswap_vsetq_lane_u64(vget_lane_u64(__s2_21, __p3_21), __rev0_21, __p1_21); \
   __ret_21 = __builtin_shufflevector(__ret_21, __ret_21, 1, 0); \
   __ret_21; \
 })
@@ -51483,7 +43595,7 @@
   int64x1_t __s2_31 = __p2_31; \
   int64x2_t __rev0_31;  __rev0_31 = __builtin_shufflevector(__s0_31, __s0_31, 1, 0); \
   int64x2_t __ret_31; \
-  __ret_31 = __noswap_vsetq_lane_s64(__noswap_vget_lane_s64(__s2_31, __p3_31), __rev0_31, __p1_31); \
+  __ret_31 = __noswap_vsetq_lane_s64(vget_lane_s64(__s2_31, __p3_31), __rev0_31, __p1_31); \
   __ret_31 = __builtin_shufflevector(__ret_31, __ret_31, 1, 0); \
   __ret_31; \
 })
@@ -51594,7 +43706,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcopy_lane_u64(__p0_42, __p1_42, __p2_42, __p3_42) __extension__ ({ \
   uint64x1_t __s0_42 = __p0_42; \
   uint64x1_t __s2_42 = __p2_42; \
@@ -51602,653 +43713,596 @@
   __ret_42 = vset_lane_u64(vget_lane_u64(__s2_42, __p3_42), __s0_42, __p1_42); \
   __ret_42; \
 })
-#else
-#define vcopy_lane_u64(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
-  uint64x1_t __s0_43 = __p0_43; \
-  uint64x1_t __s2_43 = __p2_43; \
-  uint64x1_t __ret_43; \
-  __ret_43 = __noswap_vset_lane_u64(__noswap_vget_lane_u64(__s2_43, __p3_43), __s0_43, __p1_43); \
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u16(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
+  uint16x4_t __s0_43 = __p0_43; \
+  uint16x4_t __s2_43 = __p2_43; \
+  uint16x4_t __ret_43; \
+  __ret_43 = vset_lane_u16(vget_lane_u16(__s2_43, __p3_43), __s0_43, __p1_43); \
   __ret_43; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
+#else
 #define vcopy_lane_u16(__p0_44, __p1_44, __p2_44, __p3_44) __extension__ ({ \
   uint16x4_t __s0_44 = __p0_44; \
   uint16x4_t __s2_44 = __p2_44; \
+  uint16x4_t __rev0_44;  __rev0_44 = __builtin_shufflevector(__s0_44, __s0_44, 3, 2, 1, 0); \
+  uint16x4_t __rev2_44;  __rev2_44 = __builtin_shufflevector(__s2_44, __s2_44, 3, 2, 1, 0); \
   uint16x4_t __ret_44; \
-  __ret_44 = vset_lane_u16(vget_lane_u16(__s2_44, __p3_44), __s0_44, __p1_44); \
+  __ret_44 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_44, __p3_44), __rev0_44, __p1_44); \
+  __ret_44 = __builtin_shufflevector(__ret_44, __ret_44, 3, 2, 1, 0); \
   __ret_44; \
 })
-#else
-#define vcopy_lane_u16(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
-  uint16x4_t __s0_45 = __p0_45; \
-  uint16x4_t __s2_45 = __p2_45; \
-  uint16x4_t __rev0_45;  __rev0_45 = __builtin_shufflevector(__s0_45, __s0_45, 3, 2, 1, 0); \
-  uint16x4_t __rev2_45;  __rev2_45 = __builtin_shufflevector(__s2_45, __s2_45, 3, 2, 1, 0); \
-  uint16x4_t __ret_45; \
-  __ret_45 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_45, __p3_45), __rev0_45, __p1_45); \
-  __ret_45 = __builtin_shufflevector(__ret_45, __ret_45, 3, 2, 1, 0); \
-  __ret_45; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s8(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
+  int8x8_t __s0_45 = __p0_45; \
+  int8x8_t __s2_45 = __p2_45; \
+  int8x8_t __ret_45; \
+  __ret_45 = vset_lane_s8(vget_lane_s8(__s2_45, __p3_45), __s0_45, __p1_45); \
+  __ret_45; \
+})
+#else
 #define vcopy_lane_s8(__p0_46, __p1_46, __p2_46, __p3_46) __extension__ ({ \
   int8x8_t __s0_46 = __p0_46; \
   int8x8_t __s2_46 = __p2_46; \
+  int8x8_t __rev0_46;  __rev0_46 = __builtin_shufflevector(__s0_46, __s0_46, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev2_46;  __rev2_46 = __builtin_shufflevector(__s2_46, __s2_46, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x8_t __ret_46; \
-  __ret_46 = vset_lane_s8(vget_lane_s8(__s2_46, __p3_46), __s0_46, __p1_46); \
+  __ret_46 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_46, __p3_46), __rev0_46, __p1_46); \
+  __ret_46 = __builtin_shufflevector(__ret_46, __ret_46, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_46; \
 })
-#else
-#define vcopy_lane_s8(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
-  int8x8_t __s0_47 = __p0_47; \
-  int8x8_t __s2_47 = __p2_47; \
-  int8x8_t __rev0_47;  __rev0_47 = __builtin_shufflevector(__s0_47, __s0_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2_47;  __rev2_47 = __builtin_shufflevector(__s2_47, __s2_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_47; \
-  __ret_47 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_47, __p3_47), __rev0_47, __p1_47); \
-  __ret_47 = __builtin_shufflevector(__ret_47, __ret_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_47; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_f32(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
+  float32x2_t __s0_47 = __p0_47; \
+  float32x2_t __s2_47 = __p2_47; \
+  float32x2_t __ret_47; \
+  __ret_47 = vset_lane_f32(vget_lane_f32(__s2_47, __p3_47), __s0_47, __p1_47); \
+  __ret_47; \
+})
+#else
 #define vcopy_lane_f32(__p0_48, __p1_48, __p2_48, __p3_48) __extension__ ({ \
   float32x2_t __s0_48 = __p0_48; \
   float32x2_t __s2_48 = __p2_48; \
+  float32x2_t __rev0_48;  __rev0_48 = __builtin_shufflevector(__s0_48, __s0_48, 1, 0); \
+  float32x2_t __rev2_48;  __rev2_48 = __builtin_shufflevector(__s2_48, __s2_48, 1, 0); \
   float32x2_t __ret_48; \
-  __ret_48 = vset_lane_f32(vget_lane_f32(__s2_48, __p3_48), __s0_48, __p1_48); \
+  __ret_48 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_48, __p3_48), __rev0_48, __p1_48); \
+  __ret_48 = __builtin_shufflevector(__ret_48, __ret_48, 1, 0); \
   __ret_48; \
 })
-#else
-#define vcopy_lane_f32(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
-  float32x2_t __s0_49 = __p0_49; \
-  float32x2_t __s2_49 = __p2_49; \
-  float32x2_t __rev0_49;  __rev0_49 = __builtin_shufflevector(__s0_49, __s0_49, 1, 0); \
-  float32x2_t __rev2_49;  __rev2_49 = __builtin_shufflevector(__s2_49, __s2_49, 1, 0); \
-  float32x2_t __ret_49; \
-  __ret_49 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_49, __p3_49), __rev0_49, __p1_49); \
-  __ret_49 = __builtin_shufflevector(__ret_49, __ret_49, 1, 0); \
-  __ret_49; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s32(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
+  int32x2_t __s0_49 = __p0_49; \
+  int32x2_t __s2_49 = __p2_49; \
+  int32x2_t __ret_49; \
+  __ret_49 = vset_lane_s32(vget_lane_s32(__s2_49, __p3_49), __s0_49, __p1_49); \
+  __ret_49; \
+})
+#else
 #define vcopy_lane_s32(__p0_50, __p1_50, __p2_50, __p3_50) __extension__ ({ \
   int32x2_t __s0_50 = __p0_50; \
   int32x2_t __s2_50 = __p2_50; \
+  int32x2_t __rev0_50;  __rev0_50 = __builtin_shufflevector(__s0_50, __s0_50, 1, 0); \
+  int32x2_t __rev2_50;  __rev2_50 = __builtin_shufflevector(__s2_50, __s2_50, 1, 0); \
   int32x2_t __ret_50; \
-  __ret_50 = vset_lane_s32(vget_lane_s32(__s2_50, __p3_50), __s0_50, __p1_50); \
+  __ret_50 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_50, __p3_50), __rev0_50, __p1_50); \
+  __ret_50 = __builtin_shufflevector(__ret_50, __ret_50, 1, 0); \
   __ret_50; \
 })
-#else
-#define vcopy_lane_s32(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
-  int32x2_t __s0_51 = __p0_51; \
-  int32x2_t __s2_51 = __p2_51; \
-  int32x2_t __rev0_51;  __rev0_51 = __builtin_shufflevector(__s0_51, __s0_51, 1, 0); \
-  int32x2_t __rev2_51;  __rev2_51 = __builtin_shufflevector(__s2_51, __s2_51, 1, 0); \
-  int32x2_t __ret_51; \
-  __ret_51 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_51, __p3_51), __rev0_51, __p1_51); \
-  __ret_51 = __builtin_shufflevector(__ret_51, __ret_51, 1, 0); \
-  __ret_51; \
-})
 #endif
 
+#define vcopy_lane_s64(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
+  int64x1_t __s0_51 = __p0_51; \
+  int64x1_t __s2_51 = __p2_51; \
+  int64x1_t __ret_51; \
+  __ret_51 = vset_lane_s64(vget_lane_s64(__s2_51, __p3_51), __s0_51, __p1_51); \
+  __ret_51; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s64(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
-  int64x1_t __s0_52 = __p0_52; \
-  int64x1_t __s2_52 = __p2_52; \
-  int64x1_t __ret_52; \
-  __ret_52 = vset_lane_s64(vget_lane_s64(__s2_52, __p3_52), __s0_52, __p1_52); \
+#define vcopy_lane_s16(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
+  int16x4_t __s0_52 = __p0_52; \
+  int16x4_t __s2_52 = __p2_52; \
+  int16x4_t __ret_52; \
+  __ret_52 = vset_lane_s16(vget_lane_s16(__s2_52, __p3_52), __s0_52, __p1_52); \
   __ret_52; \
 })
 #else
-#define vcopy_lane_s64(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
-  int64x1_t __s0_53 = __p0_53; \
-  int64x1_t __s2_53 = __p2_53; \
-  int64x1_t __ret_53; \
-  __ret_53 = __noswap_vset_lane_s64(__noswap_vget_lane_s64(__s2_53, __p3_53), __s0_53, __p1_53); \
+#define vcopy_lane_s16(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
+  int16x4_t __s0_53 = __p0_53; \
+  int16x4_t __s2_53 = __p2_53; \
+  int16x4_t __rev0_53;  __rev0_53 = __builtin_shufflevector(__s0_53, __s0_53, 3, 2, 1, 0); \
+  int16x4_t __rev2_53;  __rev2_53 = __builtin_shufflevector(__s2_53, __s2_53, 3, 2, 1, 0); \
+  int16x4_t __ret_53; \
+  __ret_53 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_53, __p3_53), __rev0_53, __p1_53); \
+  __ret_53 = __builtin_shufflevector(__ret_53, __ret_53, 3, 2, 1, 0); \
   __ret_53; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s16(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
-  int16x4_t __s0_54 = __p0_54; \
-  int16x4_t __s2_54 = __p2_54; \
-  int16x4_t __ret_54; \
-  __ret_54 = vset_lane_s16(vget_lane_s16(__s2_54, __p3_54), __s0_54, __p1_54); \
+#define vcopyq_laneq_p8(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
+  poly8x16_t __s0_54 = __p0_54; \
+  poly8x16_t __s2_54 = __p2_54; \
+  poly8x16_t __ret_54; \
+  __ret_54 = vsetq_lane_p8(vgetq_lane_p8(__s2_54, __p3_54), __s0_54, __p1_54); \
   __ret_54; \
 })
 #else
-#define vcopy_lane_s16(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
-  int16x4_t __s0_55 = __p0_55; \
-  int16x4_t __s2_55 = __p2_55; \
-  int16x4_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 3, 2, 1, 0); \
-  int16x4_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 3, 2, 1, 0); \
-  int16x4_t __ret_55; \
-  __ret_55 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_55, __p3_55), __rev0_55, __p1_55); \
-  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 3, 2, 1, 0); \
+#define vcopyq_laneq_p8(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
+  poly8x16_t __s0_55 = __p0_55; \
+  poly8x16_t __s2_55 = __p2_55; \
+  poly8x16_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret_55; \
+  __ret_55 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_55, __p3_55), __rev0_55, __p1_55); \
+  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_55; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p8(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
-  poly8x16_t __s0_56 = __p0_56; \
-  poly8x16_t __s2_56 = __p2_56; \
-  poly8x16_t __ret_56; \
-  __ret_56 = vsetq_lane_p8(vgetq_lane_p8(__s2_56, __p3_56), __s0_56, __p1_56); \
+#define vcopyq_laneq_p16(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
+  poly16x8_t __s0_56 = __p0_56; \
+  poly16x8_t __s2_56 = __p2_56; \
+  poly16x8_t __ret_56; \
+  __ret_56 = vsetq_lane_p16(vgetq_lane_p16(__s2_56, __p3_56), __s0_56, __p1_56); \
   __ret_56; \
 })
 #else
-#define vcopyq_laneq_p8(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
-  poly8x16_t __s0_57 = __p0_57; \
-  poly8x16_t __s2_57 = __p2_57; \
-  poly8x16_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret_57; \
-  __ret_57 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_57, __p3_57), __rev0_57, __p1_57); \
-  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_p16(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
+  poly16x8_t __s0_57 = __p0_57; \
+  poly16x8_t __s2_57 = __p2_57; \
+  poly16x8_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret_57; \
+  __ret_57 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_57, __p3_57), __rev0_57, __p1_57); \
+  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_57; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p16(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
-  poly16x8_t __s0_58 = __p0_58; \
-  poly16x8_t __s2_58 = __p2_58; \
-  poly16x8_t __ret_58; \
-  __ret_58 = vsetq_lane_p16(vgetq_lane_p16(__s2_58, __p3_58), __s0_58, __p1_58); \
+#define vcopyq_laneq_u8(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
+  uint8x16_t __s0_58 = __p0_58; \
+  uint8x16_t __s2_58 = __p2_58; \
+  uint8x16_t __ret_58; \
+  __ret_58 = vsetq_lane_u8(vgetq_lane_u8(__s2_58, __p3_58), __s0_58, __p1_58); \
   __ret_58; \
 })
 #else
-#define vcopyq_laneq_p16(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
-  poly16x8_t __s0_59 = __p0_59; \
-  poly16x8_t __s2_59 = __p2_59; \
-  poly16x8_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret_59; \
-  __ret_59 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_59, __p3_59), __rev0_59, __p1_59); \
-  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_u8(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
+  uint8x16_t __s0_59 = __p0_59; \
+  uint8x16_t __s2_59 = __p2_59; \
+  uint8x16_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_59; \
+  __ret_59 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_59, __p3_59), __rev0_59, __p1_59); \
+  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_59; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u8(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
-  uint8x16_t __s0_60 = __p0_60; \
-  uint8x16_t __s2_60 = __p2_60; \
-  uint8x16_t __ret_60; \
-  __ret_60 = vsetq_lane_u8(vgetq_lane_u8(__s2_60, __p3_60), __s0_60, __p1_60); \
+#define vcopyq_laneq_u32(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
+  uint32x4_t __s0_60 = __p0_60; \
+  uint32x4_t __s2_60 = __p2_60; \
+  uint32x4_t __ret_60; \
+  __ret_60 = vsetq_lane_u32(vgetq_lane_u32(__s2_60, __p3_60), __s0_60, __p1_60); \
   __ret_60; \
 })
 #else
-#define vcopyq_laneq_u8(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
-  uint8x16_t __s0_61 = __p0_61; \
-  uint8x16_t __s2_61 = __p2_61; \
-  uint8x16_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_61; \
-  __ret_61 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_61, __p3_61), __rev0_61, __p1_61); \
-  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_u32(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
+  uint32x4_t __s0_61 = __p0_61; \
+  uint32x4_t __s2_61 = __p2_61; \
+  uint32x4_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 3, 2, 1, 0); \
+  uint32x4_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 3, 2, 1, 0); \
+  uint32x4_t __ret_61; \
+  __ret_61 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_61, __p3_61), __rev0_61, __p1_61); \
+  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 3, 2, 1, 0); \
   __ret_61; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u32(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
-  uint32x4_t __s0_62 = __p0_62; \
-  uint32x4_t __s2_62 = __p2_62; \
-  uint32x4_t __ret_62; \
-  __ret_62 = vsetq_lane_u32(vgetq_lane_u32(__s2_62, __p3_62), __s0_62, __p1_62); \
+#define vcopyq_laneq_u64(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
+  uint64x2_t __s0_62 = __p0_62; \
+  uint64x2_t __s2_62 = __p2_62; \
+  uint64x2_t __ret_62; \
+  __ret_62 = vsetq_lane_u64(vgetq_lane_u64(__s2_62, __p3_62), __s0_62, __p1_62); \
   __ret_62; \
 })
 #else
-#define vcopyq_laneq_u32(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
-  uint32x4_t __s0_63 = __p0_63; \
-  uint32x4_t __s2_63 = __p2_63; \
-  uint32x4_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 3, 2, 1, 0); \
-  uint32x4_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 3, 2, 1, 0); \
-  uint32x4_t __ret_63; \
-  __ret_63 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_63, __p3_63), __rev0_63, __p1_63); \
-  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 3, 2, 1, 0); \
+#define vcopyq_laneq_u64(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
+  uint64x2_t __s0_63 = __p0_63; \
+  uint64x2_t __s2_63 = __p2_63; \
+  uint64x2_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 1, 0); \
+  uint64x2_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 1, 0); \
+  uint64x2_t __ret_63; \
+  __ret_63 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_63, __p3_63), __rev0_63, __p1_63); \
+  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 1, 0); \
   __ret_63; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u64(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
-  uint64x2_t __s0_64 = __p0_64; \
-  uint64x2_t __s2_64 = __p2_64; \
-  uint64x2_t __ret_64; \
-  __ret_64 = vsetq_lane_u64(vgetq_lane_u64(__s2_64, __p3_64), __s0_64, __p1_64); \
+#define vcopyq_laneq_u16(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
+  uint16x8_t __s0_64 = __p0_64; \
+  uint16x8_t __s2_64 = __p2_64; \
+  uint16x8_t __ret_64; \
+  __ret_64 = vsetq_lane_u16(vgetq_lane_u16(__s2_64, __p3_64), __s0_64, __p1_64); \
   __ret_64; \
 })
 #else
-#define vcopyq_laneq_u64(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
-  uint64x2_t __s0_65 = __p0_65; \
-  uint64x2_t __s2_65 = __p2_65; \
-  uint64x2_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 1, 0); \
-  uint64x2_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 1, 0); \
-  uint64x2_t __ret_65; \
-  __ret_65 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_65, __p3_65), __rev0_65, __p1_65); \
-  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 1, 0); \
+#define vcopyq_laneq_u16(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
+  uint16x8_t __s0_65 = __p0_65; \
+  uint16x8_t __s2_65 = __p2_65; \
+  uint16x8_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_65; \
+  __ret_65 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_65, __p3_65), __rev0_65, __p1_65); \
+  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_65; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u16(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
-  uint16x8_t __s0_66 = __p0_66; \
-  uint16x8_t __s2_66 = __p2_66; \
-  uint16x8_t __ret_66; \
-  __ret_66 = vsetq_lane_u16(vgetq_lane_u16(__s2_66, __p3_66), __s0_66, __p1_66); \
+#define vcopyq_laneq_s8(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
+  int8x16_t __s0_66 = __p0_66; \
+  int8x16_t __s2_66 = __p2_66; \
+  int8x16_t __ret_66; \
+  __ret_66 = vsetq_lane_s8(vgetq_lane_s8(__s2_66, __p3_66), __s0_66, __p1_66); \
   __ret_66; \
 })
 #else
-#define vcopyq_laneq_u16(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
-  uint16x8_t __s0_67 = __p0_67; \
-  uint16x8_t __s2_67 = __p2_67; \
-  uint16x8_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_67; \
-  __ret_67 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_67, __p3_67), __rev0_67, __p1_67); \
-  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_s8(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
+  int8x16_t __s0_67 = __p0_67; \
+  int8x16_t __s2_67 = __p2_67; \
+  int8x16_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_67; \
+  __ret_67 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_67, __p3_67), __rev0_67, __p1_67); \
+  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_67; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s8(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
-  int8x16_t __s0_68 = __p0_68; \
-  int8x16_t __s2_68 = __p2_68; \
-  int8x16_t __ret_68; \
-  __ret_68 = vsetq_lane_s8(vgetq_lane_s8(__s2_68, __p3_68), __s0_68, __p1_68); \
+#define vcopyq_laneq_f32(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
+  float32x4_t __s0_68 = __p0_68; \
+  float32x4_t __s2_68 = __p2_68; \
+  float32x4_t __ret_68; \
+  __ret_68 = vsetq_lane_f32(vgetq_lane_f32(__s2_68, __p3_68), __s0_68, __p1_68); \
   __ret_68; \
 })
 #else
-#define vcopyq_laneq_s8(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
-  int8x16_t __s0_69 = __p0_69; \
-  int8x16_t __s2_69 = __p2_69; \
-  int8x16_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_69; \
-  __ret_69 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_69, __p3_69), __rev0_69, __p1_69); \
-  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_f32(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
+  float32x4_t __s0_69 = __p0_69; \
+  float32x4_t __s2_69 = __p2_69; \
+  float32x4_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 3, 2, 1, 0); \
+  float32x4_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 3, 2, 1, 0); \
+  float32x4_t __ret_69; \
+  __ret_69 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_69, __p3_69), __rev0_69, __p1_69); \
+  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 3, 2, 1, 0); \
   __ret_69; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
-  float32x4_t __s0_70 = __p0_70; \
-  float32x4_t __s2_70 = __p2_70; \
-  float32x4_t __ret_70; \
-  __ret_70 = vsetq_lane_f32(vgetq_lane_f32(__s2_70, __p3_70), __s0_70, __p1_70); \
+#define vcopyq_laneq_s32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
+  int32x4_t __s0_70 = __p0_70; \
+  int32x4_t __s2_70 = __p2_70; \
+  int32x4_t __ret_70; \
+  __ret_70 = vsetq_lane_s32(vgetq_lane_s32(__s2_70, __p3_70), __s0_70, __p1_70); \
   __ret_70; \
 })
 #else
-#define vcopyq_laneq_f32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
-  float32x4_t __s0_71 = __p0_71; \
-  float32x4_t __s2_71 = __p2_71; \
-  float32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
-  float32x4_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 3, 2, 1, 0); \
-  float32x4_t __ret_71; \
-  __ret_71 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_71, __p3_71), __rev0_71, __p1_71); \
+#define vcopyq_laneq_s32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
+  int32x4_t __s0_71 = __p0_71; \
+  int32x4_t __s2_71 = __p2_71; \
+  int32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
+  int32x4_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 3, 2, 1, 0); \
+  int32x4_t __ret_71; \
+  __ret_71 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_71, __p3_71), __rev0_71, __p1_71); \
   __ret_71 = __builtin_shufflevector(__ret_71, __ret_71, 3, 2, 1, 0); \
   __ret_71; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s32(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
-  int32x4_t __s0_72 = __p0_72; \
-  int32x4_t __s2_72 = __p2_72; \
-  int32x4_t __ret_72; \
-  __ret_72 = vsetq_lane_s32(vgetq_lane_s32(__s2_72, __p3_72), __s0_72, __p1_72); \
+#define vcopyq_laneq_s64(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
+  int64x2_t __s0_72 = __p0_72; \
+  int64x2_t __s2_72 = __p2_72; \
+  int64x2_t __ret_72; \
+  __ret_72 = vsetq_lane_s64(vgetq_lane_s64(__s2_72, __p3_72), __s0_72, __p1_72); \
   __ret_72; \
 })
 #else
-#define vcopyq_laneq_s32(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
-  int32x4_t __s0_73 = __p0_73; \
-  int32x4_t __s2_73 = __p2_73; \
-  int32x4_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 3, 2, 1, 0); \
-  int32x4_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 3, 2, 1, 0); \
-  int32x4_t __ret_73; \
-  __ret_73 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_73, __p3_73), __rev0_73, __p1_73); \
-  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 3, 2, 1, 0); \
+#define vcopyq_laneq_s64(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
+  int64x2_t __s0_73 = __p0_73; \
+  int64x2_t __s2_73 = __p2_73; \
+  int64x2_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 1, 0); \
+  int64x2_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 1, 0); \
+  int64x2_t __ret_73; \
+  __ret_73 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_73, __p3_73), __rev0_73, __p1_73); \
+  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 1, 0); \
   __ret_73; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s64(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
-  int64x2_t __s0_74 = __p0_74; \
-  int64x2_t __s2_74 = __p2_74; \
-  int64x2_t __ret_74; \
-  __ret_74 = vsetq_lane_s64(vgetq_lane_s64(__s2_74, __p3_74), __s0_74, __p1_74); \
+#define vcopyq_laneq_s16(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
+  int16x8_t __s0_74 = __p0_74; \
+  int16x8_t __s2_74 = __p2_74; \
+  int16x8_t __ret_74; \
+  __ret_74 = vsetq_lane_s16(vgetq_lane_s16(__s2_74, __p3_74), __s0_74, __p1_74); \
   __ret_74; \
 })
 #else
-#define vcopyq_laneq_s64(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
-  int64x2_t __s0_75 = __p0_75; \
-  int64x2_t __s2_75 = __p2_75; \
-  int64x2_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 1, 0); \
-  int64x2_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 1, 0); \
-  int64x2_t __ret_75; \
-  __ret_75 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_75, __p3_75), __rev0_75, __p1_75); \
-  __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 1, 0); \
+#define vcopyq_laneq_s16(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
+  int16x8_t __s0_75 = __p0_75; \
+  int16x8_t __s2_75 = __p2_75; \
+  int16x8_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_75; \
+  __ret_75 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_75, __p3_75), __rev0_75, __p1_75); \
+  __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_75; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s16(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
-  int16x8_t __s0_76 = __p0_76; \
-  int16x8_t __s2_76 = __p2_76; \
-  int16x8_t __ret_76; \
-  __ret_76 = vsetq_lane_s16(vgetq_lane_s16(__s2_76, __p3_76), __s0_76, __p1_76); \
+#define vcopy_laneq_p8(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
+  poly8x8_t __s0_76 = __p0_76; \
+  poly8x16_t __s2_76 = __p2_76; \
+  poly8x8_t __ret_76; \
+  __ret_76 = vset_lane_p8(vgetq_lane_p8(__s2_76, __p3_76), __s0_76, __p1_76); \
   __ret_76; \
 })
 #else
-#define vcopyq_laneq_s16(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
-  int16x8_t __s0_77 = __p0_77; \
-  int16x8_t __s2_77 = __p2_77; \
-  int16x8_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_77; \
-  __ret_77 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_77, __p3_77), __rev0_77, __p1_77); \
+#define vcopy_laneq_p8(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
+  poly8x8_t __s0_77 = __p0_77; \
+  poly8x16_t __s2_77 = __p2_77; \
+  poly8x8_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret_77; \
+  __ret_77 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_77, __p3_77), __rev0_77, __p1_77); \
   __ret_77 = __builtin_shufflevector(__ret_77, __ret_77, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_77; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p8(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
-  poly8x8_t __s0_78 = __p0_78; \
-  poly8x16_t __s2_78 = __p2_78; \
-  poly8x8_t __ret_78; \
-  __ret_78 = vset_lane_p8(vgetq_lane_p8(__s2_78, __p3_78), __s0_78, __p1_78); \
+#define vcopy_laneq_p16(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
+  poly16x4_t __s0_78 = __p0_78; \
+  poly16x8_t __s2_78 = __p2_78; \
+  poly16x4_t __ret_78; \
+  __ret_78 = vset_lane_p16(vgetq_lane_p16(__s2_78, __p3_78), __s0_78, __p1_78); \
   __ret_78; \
 })
 #else
-#define vcopy_laneq_p8(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
-  poly8x8_t __s0_79 = __p0_79; \
-  poly8x16_t __s2_79 = __p2_79; \
-  poly8x8_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret_79; \
-  __ret_79 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_79, __p3_79), __rev0_79, __p1_79); \
-  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_p16(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
+  poly16x4_t __s0_79 = __p0_79; \
+  poly16x8_t __s2_79 = __p2_79; \
+  poly16x4_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 3, 2, 1, 0); \
+  poly16x8_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x4_t __ret_79; \
+  __ret_79 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_79, __p3_79), __rev0_79, __p1_79); \
+  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 3, 2, 1, 0); \
   __ret_79; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p16(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
-  poly16x4_t __s0_80 = __p0_80; \
-  poly16x8_t __s2_80 = __p2_80; \
-  poly16x4_t __ret_80; \
-  __ret_80 = vset_lane_p16(vgetq_lane_p16(__s2_80, __p3_80), __s0_80, __p1_80); \
+#define vcopy_laneq_u8(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
+  uint8x8_t __s0_80 = __p0_80; \
+  uint8x16_t __s2_80 = __p2_80; \
+  uint8x8_t __ret_80; \
+  __ret_80 = vset_lane_u8(vgetq_lane_u8(__s2_80, __p3_80), __s0_80, __p1_80); \
   __ret_80; \
 })
 #else
-#define vcopy_laneq_p16(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
-  poly16x4_t __s0_81 = __p0_81; \
-  poly16x8_t __s2_81 = __p2_81; \
-  poly16x4_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 3, 2, 1, 0); \
-  poly16x8_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x4_t __ret_81; \
-  __ret_81 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_81, __p3_81), __rev0_81, __p1_81); \
-  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 3, 2, 1, 0); \
+#define vcopy_laneq_u8(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
+  uint8x8_t __s0_81 = __p0_81; \
+  uint8x16_t __s2_81 = __p2_81; \
+  uint8x8_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret_81; \
+  __ret_81 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_81, __p3_81), __rev0_81, __p1_81); \
+  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_81; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u8(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
-  uint8x8_t __s0_82 = __p0_82; \
-  uint8x16_t __s2_82 = __p2_82; \
-  uint8x8_t __ret_82; \
-  __ret_82 = vset_lane_u8(vgetq_lane_u8(__s2_82, __p3_82), __s0_82, __p1_82); \
+#define vcopy_laneq_u32(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
+  uint32x2_t __s0_82 = __p0_82; \
+  uint32x4_t __s2_82 = __p2_82; \
+  uint32x2_t __ret_82; \
+  __ret_82 = vset_lane_u32(vgetq_lane_u32(__s2_82, __p3_82), __s0_82, __p1_82); \
   __ret_82; \
 })
 #else
-#define vcopy_laneq_u8(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
-  uint8x8_t __s0_83 = __p0_83; \
-  uint8x16_t __s2_83 = __p2_83; \
-  uint8x8_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret_83; \
-  __ret_83 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_83, __p3_83), __rev0_83, __p1_83); \
-  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_u32(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
+  uint32x2_t __s0_83 = __p0_83; \
+  uint32x4_t __s2_83 = __p2_83; \
+  uint32x2_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 1, 0); \
+  uint32x4_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 3, 2, 1, 0); \
+  uint32x2_t __ret_83; \
+  __ret_83 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_83, __p3_83), __rev0_83, __p1_83); \
+  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 1, 0); \
   __ret_83; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u32(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
-  uint32x2_t __s0_84 = __p0_84; \
-  uint32x4_t __s2_84 = __p2_84; \
-  uint32x2_t __ret_84; \
-  __ret_84 = vset_lane_u32(vgetq_lane_u32(__s2_84, __p3_84), __s0_84, __p1_84); \
+#define vcopy_laneq_u64(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
+  uint64x1_t __s0_84 = __p0_84; \
+  uint64x2_t __s2_84 = __p2_84; \
+  uint64x1_t __ret_84; \
+  __ret_84 = vset_lane_u64(vgetq_lane_u64(__s2_84, __p3_84), __s0_84, __p1_84); \
   __ret_84; \
 })
 #else
-#define vcopy_laneq_u32(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
-  uint32x2_t __s0_85 = __p0_85; \
-  uint32x4_t __s2_85 = __p2_85; \
-  uint32x2_t __rev0_85;  __rev0_85 = __builtin_shufflevector(__s0_85, __s0_85, 1, 0); \
-  uint32x4_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 3, 2, 1, 0); \
-  uint32x2_t __ret_85; \
-  __ret_85 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_85, __p3_85), __rev0_85, __p1_85); \
-  __ret_85 = __builtin_shufflevector(__ret_85, __ret_85, 1, 0); \
+#define vcopy_laneq_u64(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
+  uint64x1_t __s0_85 = __p0_85; \
+  uint64x2_t __s2_85 = __p2_85; \
+  uint64x2_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 1, 0); \
+  uint64x1_t __ret_85; \
+  __ret_85 = vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_85, __p3_85), __s0_85, __p1_85); \
   __ret_85; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u64(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
-  uint64x1_t __s0_86 = __p0_86; \
-  uint64x2_t __s2_86 = __p2_86; \
-  uint64x1_t __ret_86; \
-  __ret_86 = vset_lane_u64(vgetq_lane_u64(__s2_86, __p3_86), __s0_86, __p1_86); \
+#define vcopy_laneq_u16(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
+  uint16x4_t __s0_86 = __p0_86; \
+  uint16x8_t __s2_86 = __p2_86; \
+  uint16x4_t __ret_86; \
+  __ret_86 = vset_lane_u16(vgetq_lane_u16(__s2_86, __p3_86), __s0_86, __p1_86); \
   __ret_86; \
 })
 #else
-#define vcopy_laneq_u64(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
-  uint64x1_t __s0_87 = __p0_87; \
-  uint64x2_t __s2_87 = __p2_87; \
-  uint64x2_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 1, 0); \
-  uint64x1_t __ret_87; \
-  __ret_87 = __noswap_vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_87, __p3_87), __s0_87, __p1_87); \
+#define vcopy_laneq_u16(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
+  uint16x4_t __s0_87 = __p0_87; \
+  uint16x8_t __s2_87 = __p2_87; \
+  uint16x4_t __rev0_87;  __rev0_87 = __builtin_shufflevector(__s0_87, __s0_87, 3, 2, 1, 0); \
+  uint16x8_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret_87; \
+  __ret_87 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_87, __p3_87), __rev0_87, __p1_87); \
+  __ret_87 = __builtin_shufflevector(__ret_87, __ret_87, 3, 2, 1, 0); \
   __ret_87; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u16(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
-  uint16x4_t __s0_88 = __p0_88; \
-  uint16x8_t __s2_88 = __p2_88; \
-  uint16x4_t __ret_88; \
-  __ret_88 = vset_lane_u16(vgetq_lane_u16(__s2_88, __p3_88), __s0_88, __p1_88); \
+#define vcopy_laneq_s8(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
+  int8x8_t __s0_88 = __p0_88; \
+  int8x16_t __s2_88 = __p2_88; \
+  int8x8_t __ret_88; \
+  __ret_88 = vset_lane_s8(vgetq_lane_s8(__s2_88, __p3_88), __s0_88, __p1_88); \
   __ret_88; \
 })
 #else
-#define vcopy_laneq_u16(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
-  uint16x4_t __s0_89 = __p0_89; \
-  uint16x8_t __s2_89 = __p2_89; \
-  uint16x4_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 3, 2, 1, 0); \
-  uint16x8_t __rev2_89;  __rev2_89 = __builtin_shufflevector(__s2_89, __s2_89, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret_89; \
-  __ret_89 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_89, __p3_89), __rev0_89, __p1_89); \
-  __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 3, 2, 1, 0); \
+#define vcopy_laneq_s8(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
+  int8x8_t __s0_89 = __p0_89; \
+  int8x16_t __s2_89 = __p2_89; \
+  int8x8_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_89;  __rev2_89 = __builtin_shufflevector(__s2_89, __s2_89, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret_89; \
+  __ret_89 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_89, __p3_89), __rev0_89, __p1_89); \
+  __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_89; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s8(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
-  int8x8_t __s0_90 = __p0_90; \
-  int8x16_t __s2_90 = __p2_90; \
-  int8x8_t __ret_90; \
-  __ret_90 = vset_lane_s8(vgetq_lane_s8(__s2_90, __p3_90), __s0_90, __p1_90); \
+#define vcopy_laneq_f32(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
+  float32x2_t __s0_90 = __p0_90; \
+  float32x4_t __s2_90 = __p2_90; \
+  float32x2_t __ret_90; \
+  __ret_90 = vset_lane_f32(vgetq_lane_f32(__s2_90, __p3_90), __s0_90, __p1_90); \
   __ret_90; \
 })
 #else
-#define vcopy_laneq_s8(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
-  int8x8_t __s0_91 = __p0_91; \
-  int8x16_t __s2_91 = __p2_91; \
-  int8x8_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_91; \
-  __ret_91 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_91, __p3_91), __rev0_91, __p1_91); \
-  __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_f32(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
+  float32x2_t __s0_91 = __p0_91; \
+  float32x4_t __s2_91 = __p2_91; \
+  float32x2_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 1, 0); \
+  float32x4_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 3, 2, 1, 0); \
+  float32x2_t __ret_91; \
+  __ret_91 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_91, __p3_91), __rev0_91, __p1_91); \
+  __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 1, 0); \
   __ret_91; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f32(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
-  float32x2_t __s0_92 = __p0_92; \
-  float32x4_t __s2_92 = __p2_92; \
-  float32x2_t __ret_92; \
-  __ret_92 = vset_lane_f32(vgetq_lane_f32(__s2_92, __p3_92), __s0_92, __p1_92); \
+#define vcopy_laneq_s32(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
+  int32x2_t __s0_92 = __p0_92; \
+  int32x4_t __s2_92 = __p2_92; \
+  int32x2_t __ret_92; \
+  __ret_92 = vset_lane_s32(vgetq_lane_s32(__s2_92, __p3_92), __s0_92, __p1_92); \
   __ret_92; \
 })
 #else
-#define vcopy_laneq_f32(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
-  float32x2_t __s0_93 = __p0_93; \
-  float32x4_t __s2_93 = __p2_93; \
-  float32x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
-  float32x4_t __rev2_93;  __rev2_93 = __builtin_shufflevector(__s2_93, __s2_93, 3, 2, 1, 0); \
-  float32x2_t __ret_93; \
-  __ret_93 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_93, __p3_93), __rev0_93, __p1_93); \
+#define vcopy_laneq_s32(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
+  int32x2_t __s0_93 = __p0_93; \
+  int32x4_t __s2_93 = __p2_93; \
+  int32x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
+  int32x4_t __rev2_93;  __rev2_93 = __builtin_shufflevector(__s2_93, __s2_93, 3, 2, 1, 0); \
+  int32x2_t __ret_93; \
+  __ret_93 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_93, __p3_93), __rev0_93, __p1_93); \
   __ret_93 = __builtin_shufflevector(__ret_93, __ret_93, 1, 0); \
   __ret_93; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s32(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
-  int32x2_t __s0_94 = __p0_94; \
-  int32x4_t __s2_94 = __p2_94; \
-  int32x2_t __ret_94; \
-  __ret_94 = vset_lane_s32(vgetq_lane_s32(__s2_94, __p3_94), __s0_94, __p1_94); \
+#define vcopy_laneq_s64(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
+  int64x1_t __s0_94 = __p0_94; \
+  int64x2_t __s2_94 = __p2_94; \
+  int64x1_t __ret_94; \
+  __ret_94 = vset_lane_s64(vgetq_lane_s64(__s2_94, __p3_94), __s0_94, __p1_94); \
   __ret_94; \
 })
 #else
-#define vcopy_laneq_s32(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
-  int32x2_t __s0_95 = __p0_95; \
-  int32x4_t __s2_95 = __p2_95; \
-  int32x2_t __rev0_95;  __rev0_95 = __builtin_shufflevector(__s0_95, __s0_95, 1, 0); \
-  int32x4_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 3, 2, 1, 0); \
-  int32x2_t __ret_95; \
-  __ret_95 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_95, __p3_95), __rev0_95, __p1_95); \
-  __ret_95 = __builtin_shufflevector(__ret_95, __ret_95, 1, 0); \
+#define vcopy_laneq_s64(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
+  int64x1_t __s0_95 = __p0_95; \
+  int64x2_t __s2_95 = __p2_95; \
+  int64x2_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 1, 0); \
+  int64x1_t __ret_95; \
+  __ret_95 = vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_95, __p3_95), __s0_95, __p1_95); \
   __ret_95; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s64(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
-  int64x1_t __s0_96 = __p0_96; \
-  int64x2_t __s2_96 = __p2_96; \
-  int64x1_t __ret_96; \
-  __ret_96 = vset_lane_s64(vgetq_lane_s64(__s2_96, __p3_96), __s0_96, __p1_96); \
+#define vcopy_laneq_s16(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
+  int16x4_t __s0_96 = __p0_96; \
+  int16x8_t __s2_96 = __p2_96; \
+  int16x4_t __ret_96; \
+  __ret_96 = vset_lane_s16(vgetq_lane_s16(__s2_96, __p3_96), __s0_96, __p1_96); \
   __ret_96; \
 })
 #else
-#define vcopy_laneq_s64(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
-  int64x1_t __s0_97 = __p0_97; \
-  int64x2_t __s2_97 = __p2_97; \
-  int64x2_t __rev2_97;  __rev2_97 = __builtin_shufflevector(__s2_97, __s2_97, 1, 0); \
-  int64x1_t __ret_97; \
-  __ret_97 = __noswap_vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_97, __p3_97), __s0_97, __p1_97); \
+#define vcopy_laneq_s16(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
+  int16x4_t __s0_97 = __p0_97; \
+  int16x8_t __s2_97 = __p2_97; \
+  int16x4_t __rev0_97;  __rev0_97 = __builtin_shufflevector(__s0_97, __s0_97, 3, 2, 1, 0); \
+  int16x8_t __rev2_97;  __rev2_97 = __builtin_shufflevector(__s2_97, __s2_97, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret_97; \
+  __ret_97 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_97, __p3_97), __rev0_97, __p1_97); \
+  __ret_97 = __builtin_shufflevector(__ret_97, __ret_97, 3, 2, 1, 0); \
   __ret_97; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s16(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
-  int16x4_t __s0_98 = __p0_98; \
-  int16x8_t __s2_98 = __p2_98; \
-  int16x4_t __ret_98; \
-  __ret_98 = vset_lane_s16(vgetq_lane_s16(__s2_98, __p3_98), __s0_98, __p1_98); \
-  __ret_98; \
-})
-#else
-#define vcopy_laneq_s16(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
-  int16x4_t __s0_99 = __p0_99; \
-  int16x8_t __s2_99 = __p2_99; \
-  int16x4_t __rev0_99;  __rev0_99 = __builtin_shufflevector(__s0_99, __s0_99, 3, 2, 1, 0); \
-  int16x8_t __rev2_99;  __rev2_99 = __builtin_shufflevector(__s2_99, __s2_99, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret_99; \
-  __ret_99 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_99, __p3_99), __rev0_99, __p1_99); \
-  __ret_99 = __builtin_shufflevector(__ret_99, __ret_99, 3, 2, 1, 0); \
-  __ret_99; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vcreate_p64(uint64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vcreate_p64(uint64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcreate_f64(uint64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vcreate_f64(uint64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvts_f32_s32(int32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvts_f32_s32(int32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvts_f32_u32(uint32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvts_f32_u32(uint32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
   float32x2_t __ret;
@@ -52270,34 +44324,16 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vcvtd_f64_s64(int64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vcvtd_f64_s64(int64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vcvtd_f64_u64(uint64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
   float64x2_t __ret;
@@ -52330,34 +44366,16 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
   float64x2_t __ret;
@@ -52445,38 +44463,18 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
   uint32_t __s0 = __p0; \
   float32_t __ret; \
   __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   float32_t __ret; \
   __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64x2_t __s0 = __p0; \
@@ -52513,86 +44511,36 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
   float32_t __s0 = __p0; \
   int32_t __ret; \
   __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -52611,54 +44559,24 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
   float32_t __s0 = __p0; \
   uint32_t __ret; \
   __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -52677,66 +44595,28 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvts_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvts_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -52753,48 +44633,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvts_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvts_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -52811,258 +44664,96 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtas_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtas_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtad_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtad_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtas_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtad_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtms_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtms_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtmd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtms_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtns_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtns_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtnd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtns_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtps_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtps_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtpd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtps_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvtxd_f32_f64(float64_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
   float32x2_t __ret;
@@ -53135,20 +44826,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 / __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
   float32x2_t __ret;
@@ -53234,22 +44916,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vduph_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -53284,22 +44956,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdups_lane_f32(__p0, __p1) __extension__ ({ \
   float32x2_t __s0 = __p0; \
@@ -53334,22 +44996,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vduph_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -53367,22 +45019,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
@@ -53435,22 +45077,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_f16(__p0, __p1) __extension__ ({ \
   float16x4_t __s0 = __p0; \
@@ -54173,20 +45805,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vdup_n_p64(poly64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
   poly64x2_t __ret;
@@ -54217,21 +45840,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vdup_n_f64(float64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai float64x1_t vdup_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vext_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -54239,16 +45852,6 @@
   __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -54291,7 +45894,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -54299,16 +45901,6 @@
   __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
   float64x2_t __ret;
@@ -54332,26 +45924,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-__ai float64x1_t __noswap_vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
   float64_t __s0 = __p0; \
   float64_t __s1 = __p1; \
@@ -54360,25 +45937,6 @@
   __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
   __ret; \
 })
-#else
-#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
   float32_t __s0 = __p0; \
@@ -54471,7 +46029,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -54480,25 +46037,6 @@
   __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
   __ret; \
 })
-#else
-#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#define __noswap_vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
   float32x2_t __s0 = __p0; \
@@ -54731,20 +46269,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, __p1, (float64x1_t) {__p2});
   return __ret;
 }
-#else
-__ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, __p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
   float64x2_t __ret;
@@ -54763,288 +46292,255 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, -__p1, __p2);
   return __ret;
 }
-#else
-__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, __p2);
-  return __ret;
-}
-#endif
-
+#define vfmsd_lane_f64(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
+  float64_t __s0_98 = __p0_98; \
+  float64_t __s1_98 = __p1_98; \
+  float64x1_t __s2_98 = __p2_98; \
+  float64_t __ret_98; \
+  __ret_98 = vfmad_lane_f64(__s0_98, -__s1_98, __s2_98, __p3_98); \
+  __ret_98; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vfmsd_lane_f64(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
-  float64_t __s0_100 = __p0_100; \
-  float64_t __s1_100 = __p1_100; \
-  float64x1_t __s2_100 = __p2_100; \
-  float64_t __ret_100; \
-  __ret_100 = vfmad_lane_f64(__s0_100, -__s1_100, __s2_100, __p3_100); \
+#define vfmss_lane_f32(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
+  float32_t __s0_99 = __p0_99; \
+  float32_t __s1_99 = __p1_99; \
+  float32x2_t __s2_99 = __p2_99; \
+  float32_t __ret_99; \
+  __ret_99 = vfmas_lane_f32(__s0_99, -__s1_99, __s2_99, __p3_99); \
+  __ret_99; \
+})
+#else
+#define vfmss_lane_f32(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
+  float32_t __s0_100 = __p0_100; \
+  float32_t __s1_100 = __p1_100; \
+  float32x2_t __s2_100 = __p2_100; \
+  float32x2_t __rev2_100;  __rev2_100 = __builtin_shufflevector(__s2_100, __s2_100, 1, 0); \
+  float32_t __ret_100; \
+  __ret_100 = __noswap_vfmas_lane_f32(__s0_100, -__s1_100, __rev2_100, __p3_100); \
   __ret_100; \
 })
-#else
-#define vfmsd_lane_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
-  float64_t __s0_101 = __p0_101; \
-  float64_t __s1_101 = __p1_101; \
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
+  float64x2_t __s0_101 = __p0_101; \
+  float64x2_t __s1_101 = __p1_101; \
   float64x1_t __s2_101 = __p2_101; \
-  float64_t __ret_101; \
-  __ret_101 = __noswap_vfmad_lane_f64(__s0_101, -__s1_101, __s2_101, __p3_101); \
+  float64x2_t __ret_101; \
+  __ret_101 = vfmaq_lane_f64(__s0_101, -__s1_101, __s2_101, __p3_101); \
   __ret_101; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmss_lane_f32(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
-  float32_t __s0_102 = __p0_102; \
-  float32_t __s1_102 = __p1_102; \
-  float32x2_t __s2_102 = __p2_102; \
-  float32_t __ret_102; \
-  __ret_102 = vfmas_lane_f32(__s0_102, -__s1_102, __s2_102, __p3_102); \
+#else
+#define vfmsq_lane_f64(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
+  float64x2_t __s0_102 = __p0_102; \
+  float64x2_t __s1_102 = __p1_102; \
+  float64x1_t __s2_102 = __p2_102; \
+  float64x2_t __rev0_102;  __rev0_102 = __builtin_shufflevector(__s0_102, __s0_102, 1, 0); \
+  float64x2_t __rev1_102;  __rev1_102 = __builtin_shufflevector(__s1_102, __s1_102, 1, 0); \
+  float64x2_t __ret_102; \
+  __ret_102 = __noswap_vfmaq_lane_f64(__rev0_102, -__rev1_102, __s2_102, __p3_102); \
+  __ret_102 = __builtin_shufflevector(__ret_102, __ret_102, 1, 0); \
   __ret_102; \
 })
-#else
-#define vfmss_lane_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
-  float32_t __s0_103 = __p0_103; \
-  float32_t __s1_103 = __p1_103; \
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
+  float32x4_t __s0_103 = __p0_103; \
+  float32x4_t __s1_103 = __p1_103; \
   float32x2_t __s2_103 = __p2_103; \
-  float32x2_t __rev2_103;  __rev2_103 = __builtin_shufflevector(__s2_103, __s2_103, 1, 0); \
-  float32_t __ret_103; \
-  __ret_103 = __noswap_vfmas_lane_f32(__s0_103, -__s1_103, __rev2_103, __p3_103); \
+  float32x4_t __ret_103; \
+  __ret_103 = vfmaq_lane_f32(__s0_103, -__s1_103, __s2_103, __p3_103); \
   __ret_103; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f64(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
-  float64x2_t __s0_104 = __p0_104; \
-  float64x2_t __s1_104 = __p1_104; \
-  float64x1_t __s2_104 = __p2_104; \
-  float64x2_t __ret_104; \
-  __ret_104 = vfmaq_lane_f64(__s0_104, -__s1_104, __s2_104, __p3_104); \
+#else
+#define vfmsq_lane_f32(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
+  float32x4_t __s0_104 = __p0_104; \
+  float32x4_t __s1_104 = __p1_104; \
+  float32x2_t __s2_104 = __p2_104; \
+  float32x4_t __rev0_104;  __rev0_104 = __builtin_shufflevector(__s0_104, __s0_104, 3, 2, 1, 0); \
+  float32x4_t __rev1_104;  __rev1_104 = __builtin_shufflevector(__s1_104, __s1_104, 3, 2, 1, 0); \
+  float32x2_t __rev2_104;  __rev2_104 = __builtin_shufflevector(__s2_104, __s2_104, 1, 0); \
+  float32x4_t __ret_104; \
+  __ret_104 = __noswap_vfmaq_lane_f32(__rev0_104, -__rev1_104, __rev2_104, __p3_104); \
+  __ret_104 = __builtin_shufflevector(__ret_104, __ret_104, 3, 2, 1, 0); \
   __ret_104; \
 })
-#else
-#define vfmsq_lane_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
-  float64x2_t __s0_105 = __p0_105; \
-  float64x2_t __s1_105 = __p1_105; \
-  float64x1_t __s2_105 = __p2_105; \
-  float64x2_t __rev0_105;  __rev0_105 = __builtin_shufflevector(__s0_105, __s0_105, 1, 0); \
-  float64x2_t __rev1_105;  __rev1_105 = __builtin_shufflevector(__s1_105, __s1_105, 1, 0); \
-  float64x2_t __ret_105; \
-  __ret_105 = __noswap_vfmaq_lane_f64(__rev0_105, -__rev1_105, __s2_105, __p3_105); \
-  __ret_105 = __builtin_shufflevector(__ret_105, __ret_105, 1, 0); \
-  __ret_105; \
-})
 #endif
 
+#define vfms_lane_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
+  float64x1_t __s0_105 = __p0_105; \
+  float64x1_t __s1_105 = __p1_105; \
+  float64x1_t __s2_105 = __p2_105; \
+  float64x1_t __ret_105; \
+  __ret_105 = vfma_lane_f64(__s0_105, -__s1_105, __s2_105, __p3_105); \
+  __ret_105; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
-  float32x4_t __s0_106 = __p0_106; \
-  float32x4_t __s1_106 = __p1_106; \
+#define vfms_lane_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
+  float32x2_t __s0_106 = __p0_106; \
+  float32x2_t __s1_106 = __p1_106; \
   float32x2_t __s2_106 = __p2_106; \
-  float32x4_t __ret_106; \
-  __ret_106 = vfmaq_lane_f32(__s0_106, -__s1_106, __s2_106, __p3_106); \
+  float32x2_t __ret_106; \
+  __ret_106 = vfma_lane_f32(__s0_106, -__s1_106, __s2_106, __p3_106); \
   __ret_106; \
 })
 #else
-#define vfmsq_lane_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
-  float32x4_t __s0_107 = __p0_107; \
-  float32x4_t __s1_107 = __p1_107; \
+#define vfms_lane_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
+  float32x2_t __s0_107 = __p0_107; \
+  float32x2_t __s1_107 = __p1_107; \
   float32x2_t __s2_107 = __p2_107; \
-  float32x4_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 3, 2, 1, 0); \
-  float32x4_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 3, 2, 1, 0); \
+  float32x2_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 1, 0); \
+  float32x2_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 1, 0); \
   float32x2_t __rev2_107;  __rev2_107 = __builtin_shufflevector(__s2_107, __s2_107, 1, 0); \
-  float32x4_t __ret_107; \
-  __ret_107 = __noswap_vfmaq_lane_f32(__rev0_107, -__rev1_107, __rev2_107, __p3_107); \
-  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 3, 2, 1, 0); \
+  float32x2_t __ret_107; \
+  __ret_107 = __noswap_vfma_lane_f32(__rev0_107, -__rev1_107, __rev2_107, __p3_107); \
+  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 1, 0); \
   __ret_107; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
-  float64x1_t __s0_108 = __p0_108; \
-  float64x1_t __s1_108 = __p1_108; \
-  float64x1_t __s2_108 = __p2_108; \
-  float64x1_t __ret_108; \
-  __ret_108 = vfma_lane_f64(__s0_108, -__s1_108, __s2_108, __p3_108); \
+#define vfmsd_laneq_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
+  float64_t __s0_108 = __p0_108; \
+  float64_t __s1_108 = __p1_108; \
+  float64x2_t __s2_108 = __p2_108; \
+  float64_t __ret_108; \
+  __ret_108 = vfmad_laneq_f64(__s0_108, -__s1_108, __s2_108, __p3_108); \
   __ret_108; \
 })
 #else
-#define vfms_lane_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
-  float64x1_t __s0_109 = __p0_109; \
-  float64x1_t __s1_109 = __p1_109; \
-  float64x1_t __s2_109 = __p2_109; \
-  float64x1_t __ret_109; \
-  __ret_109 = __noswap_vfma_lane_f64(__s0_109, -__s1_109, __s2_109, __p3_109); \
+#define vfmsd_laneq_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
+  float64_t __s0_109 = __p0_109; \
+  float64_t __s1_109 = __p1_109; \
+  float64x2_t __s2_109 = __p2_109; \
+  float64x2_t __rev2_109;  __rev2_109 = __builtin_shufflevector(__s2_109, __s2_109, 1, 0); \
+  float64_t __ret_109; \
+  __ret_109 = __noswap_vfmad_laneq_f64(__s0_109, -__s1_109, __rev2_109, __p3_109); \
   __ret_109; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
-  float32x2_t __s0_110 = __p0_110; \
-  float32x2_t __s1_110 = __p1_110; \
-  float32x2_t __s2_110 = __p2_110; \
-  float32x2_t __ret_110; \
-  __ret_110 = vfma_lane_f32(__s0_110, -__s1_110, __s2_110, __p3_110); \
+#define vfmss_laneq_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
+  float32_t __s0_110 = __p0_110; \
+  float32_t __s1_110 = __p1_110; \
+  float32x4_t __s2_110 = __p2_110; \
+  float32_t __ret_110; \
+  __ret_110 = vfmas_laneq_f32(__s0_110, -__s1_110, __s2_110, __p3_110); \
   __ret_110; \
 })
 #else
-#define vfms_lane_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
-  float32x2_t __s0_111 = __p0_111; \
-  float32x2_t __s1_111 = __p1_111; \
-  float32x2_t __s2_111 = __p2_111; \
-  float32x2_t __rev0_111;  __rev0_111 = __builtin_shufflevector(__s0_111, __s0_111, 1, 0); \
-  float32x2_t __rev1_111;  __rev1_111 = __builtin_shufflevector(__s1_111, __s1_111, 1, 0); \
-  float32x2_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 1, 0); \
-  float32x2_t __ret_111; \
-  __ret_111 = __noswap_vfma_lane_f32(__rev0_111, -__rev1_111, __rev2_111, __p3_111); \
-  __ret_111 = __builtin_shufflevector(__ret_111, __ret_111, 1, 0); \
+#define vfmss_laneq_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
+  float32_t __s0_111 = __p0_111; \
+  float32_t __s1_111 = __p1_111; \
+  float32x4_t __s2_111 = __p2_111; \
+  float32x4_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 3, 2, 1, 0); \
+  float32_t __ret_111; \
+  __ret_111 = __noswap_vfmas_laneq_f32(__s0_111, -__s1_111, __rev2_111, __p3_111); \
   __ret_111; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsd_laneq_f64(__p0_112, __p1_112, __p2_112, __p3_112) __extension__ ({ \
-  float64_t __s0_112 = __p0_112; \
-  float64_t __s1_112 = __p1_112; \
+#define vfmsq_laneq_f64(__p0_112, __p1_112, __p2_112, __p3_112) __extension__ ({ \
+  float64x2_t __s0_112 = __p0_112; \
+  float64x2_t __s1_112 = __p1_112; \
   float64x2_t __s2_112 = __p2_112; \
-  float64_t __ret_112; \
-  __ret_112 = vfmad_laneq_f64(__s0_112, -__s1_112, __s2_112, __p3_112); \
+  float64x2_t __ret_112; \
+  __ret_112 = vfmaq_laneq_f64(__s0_112, -__s1_112, __s2_112, __p3_112); \
   __ret_112; \
 })
 #else
-#define vfmsd_laneq_f64(__p0_113, __p1_113, __p2_113, __p3_113) __extension__ ({ \
-  float64_t __s0_113 = __p0_113; \
-  float64_t __s1_113 = __p1_113; \
+#define vfmsq_laneq_f64(__p0_113, __p1_113, __p2_113, __p3_113) __extension__ ({ \
+  float64x2_t __s0_113 = __p0_113; \
+  float64x2_t __s1_113 = __p1_113; \
   float64x2_t __s2_113 = __p2_113; \
+  float64x2_t __rev0_113;  __rev0_113 = __builtin_shufflevector(__s0_113, __s0_113, 1, 0); \
+  float64x2_t __rev1_113;  __rev1_113 = __builtin_shufflevector(__s1_113, __s1_113, 1, 0); \
   float64x2_t __rev2_113;  __rev2_113 = __builtin_shufflevector(__s2_113, __s2_113, 1, 0); \
-  float64_t __ret_113; \
-  __ret_113 = __noswap_vfmad_laneq_f64(__s0_113, -__s1_113, __rev2_113, __p3_113); \
+  float64x2_t __ret_113; \
+  __ret_113 = __noswap_vfmaq_laneq_f64(__rev0_113, -__rev1_113, __rev2_113, __p3_113); \
+  __ret_113 = __builtin_shufflevector(__ret_113, __ret_113, 1, 0); \
   __ret_113; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmss_laneq_f32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
-  float32_t __s0_114 = __p0_114; \
-  float32_t __s1_114 = __p1_114; \
+#define vfmsq_laneq_f32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
+  float32x4_t __s0_114 = __p0_114; \
+  float32x4_t __s1_114 = __p1_114; \
   float32x4_t __s2_114 = __p2_114; \
-  float32_t __ret_114; \
-  __ret_114 = vfmas_laneq_f32(__s0_114, -__s1_114, __s2_114, __p3_114); \
+  float32x4_t __ret_114; \
+  __ret_114 = vfmaq_laneq_f32(__s0_114, -__s1_114, __s2_114, __p3_114); \
   __ret_114; \
 })
 #else
-#define vfmss_laneq_f32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
-  float32_t __s0_115 = __p0_115; \
-  float32_t __s1_115 = __p1_115; \
+#define vfmsq_laneq_f32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
+  float32x4_t __s0_115 = __p0_115; \
+  float32x4_t __s1_115 = __p1_115; \
   float32x4_t __s2_115 = __p2_115; \
+  float32x4_t __rev0_115;  __rev0_115 = __builtin_shufflevector(__s0_115, __s0_115, 3, 2, 1, 0); \
+  float32x4_t __rev1_115;  __rev1_115 = __builtin_shufflevector(__s1_115, __s1_115, 3, 2, 1, 0); \
   float32x4_t __rev2_115;  __rev2_115 = __builtin_shufflevector(__s2_115, __s2_115, 3, 2, 1, 0); \
-  float32_t __ret_115; \
-  __ret_115 = __noswap_vfmas_laneq_f32(__s0_115, -__s1_115, __rev2_115, __p3_115); \
+  float32x4_t __ret_115; \
+  __ret_115 = __noswap_vfmaq_laneq_f32(__rev0_115, -__rev1_115, __rev2_115, __p3_115); \
+  __ret_115 = __builtin_shufflevector(__ret_115, __ret_115, 3, 2, 1, 0); \
   __ret_115; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f64(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
-  float64x2_t __s0_116 = __p0_116; \
-  float64x2_t __s1_116 = __p1_116; \
+#define vfms_laneq_f64(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
+  float64x1_t __s0_116 = __p0_116; \
+  float64x1_t __s1_116 = __p1_116; \
   float64x2_t __s2_116 = __p2_116; \
-  float64x2_t __ret_116; \
-  __ret_116 = vfmaq_laneq_f64(__s0_116, -__s1_116, __s2_116, __p3_116); \
+  float64x1_t __ret_116; \
+  __ret_116 = vfma_laneq_f64(__s0_116, -__s1_116, __s2_116, __p3_116); \
   __ret_116; \
 })
 #else
-#define vfmsq_laneq_f64(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
-  float64x2_t __s0_117 = __p0_117; \
-  float64x2_t __s1_117 = __p1_117; \
+#define vfms_laneq_f64(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
+  float64x1_t __s0_117 = __p0_117; \
+  float64x1_t __s1_117 = __p1_117; \
   float64x2_t __s2_117 = __p2_117; \
-  float64x2_t __rev0_117;  __rev0_117 = __builtin_shufflevector(__s0_117, __s0_117, 1, 0); \
-  float64x2_t __rev1_117;  __rev1_117 = __builtin_shufflevector(__s1_117, __s1_117, 1, 0); \
   float64x2_t __rev2_117;  __rev2_117 = __builtin_shufflevector(__s2_117, __s2_117, 1, 0); \
-  float64x2_t __ret_117; \
-  __ret_117 = __noswap_vfmaq_laneq_f64(__rev0_117, -__rev1_117, __rev2_117, __p3_117); \
-  __ret_117 = __builtin_shufflevector(__ret_117, __ret_117, 1, 0); \
+  float64x1_t __ret_117; \
+  __ret_117 = __noswap_vfma_laneq_f64(__s0_117, -__s1_117, __rev2_117, __p3_117); \
   __ret_117; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
-  float32x4_t __s0_118 = __p0_118; \
-  float32x4_t __s1_118 = __p1_118; \
+#define vfms_laneq_f32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
+  float32x2_t __s0_118 = __p0_118; \
+  float32x2_t __s1_118 = __p1_118; \
   float32x4_t __s2_118 = __p2_118; \
-  float32x4_t __ret_118; \
-  __ret_118 = vfmaq_laneq_f32(__s0_118, -__s1_118, __s2_118, __p3_118); \
+  float32x2_t __ret_118; \
+  __ret_118 = vfma_laneq_f32(__s0_118, -__s1_118, __s2_118, __p3_118); \
   __ret_118; \
 })
 #else
-#define vfmsq_laneq_f32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
-  float32x4_t __s0_119 = __p0_119; \
-  float32x4_t __s1_119 = __p1_119; \
+#define vfms_laneq_f32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
+  float32x2_t __s0_119 = __p0_119; \
+  float32x2_t __s1_119 = __p1_119; \
   float32x4_t __s2_119 = __p2_119; \
-  float32x4_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 3, 2, 1, 0); \
-  float32x4_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 3, 2, 1, 0); \
+  float32x2_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 1, 0); \
+  float32x2_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 1, 0); \
   float32x4_t __rev2_119;  __rev2_119 = __builtin_shufflevector(__s2_119, __s2_119, 3, 2, 1, 0); \
-  float32x4_t __ret_119; \
-  __ret_119 = __noswap_vfmaq_laneq_f32(__rev0_119, -__rev1_119, __rev2_119, __p3_119); \
-  __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 3, 2, 1, 0); \
+  float32x2_t __ret_119; \
+  __ret_119 = __noswap_vfma_laneq_f32(__rev0_119, -__rev1_119, __rev2_119, __p3_119); \
+  __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 1, 0); \
   __ret_119; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f64(__p0_120, __p1_120, __p2_120, __p3_120) __extension__ ({ \
-  float64x1_t __s0_120 = __p0_120; \
-  float64x1_t __s1_120 = __p1_120; \
-  float64x2_t __s2_120 = __p2_120; \
-  float64x1_t __ret_120; \
-  __ret_120 = vfma_laneq_f64(__s0_120, -__s1_120, __s2_120, __p3_120); \
-  __ret_120; \
-})
-#else
-#define vfms_laneq_f64(__p0_121, __p1_121, __p2_121, __p3_121) __extension__ ({ \
-  float64x1_t __s0_121 = __p0_121; \
-  float64x1_t __s1_121 = __p1_121; \
-  float64x2_t __s2_121 = __p2_121; \
-  float64x2_t __rev2_121;  __rev2_121 = __builtin_shufflevector(__s2_121, __s2_121, 1, 0); \
-  float64x1_t __ret_121; \
-  __ret_121 = __noswap_vfma_laneq_f64(__s0_121, -__s1_121, __rev2_121, __p3_121); \
-  __ret_121; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f32(__p0_122, __p1_122, __p2_122, __p3_122) __extension__ ({ \
-  float32x2_t __s0_122 = __p0_122; \
-  float32x2_t __s1_122 = __p1_122; \
-  float32x4_t __s2_122 = __p2_122; \
-  float32x2_t __ret_122; \
-  __ret_122 = vfma_laneq_f32(__s0_122, -__s1_122, __s2_122, __p3_122); \
-  __ret_122; \
-})
-#else
-#define vfms_laneq_f32(__p0_123, __p1_123, __p2_123, __p3_123) __extension__ ({ \
-  float32x2_t __s0_123 = __p0_123; \
-  float32x2_t __s1_123 = __p1_123; \
-  float32x4_t __s2_123 = __p2_123; \
-  float32x2_t __rev0_123;  __rev0_123 = __builtin_shufflevector(__s0_123, __s0_123, 1, 0); \
-  float32x2_t __rev1_123;  __rev1_123 = __builtin_shufflevector(__s1_123, __s1_123, 1, 0); \
-  float32x4_t __rev2_123;  __rev2_123 = __builtin_shufflevector(__s2_123, __s2_123, 3, 2, 1, 0); \
-  float32x2_t __ret_123; \
-  __ret_123 = __noswap_vfma_laneq_f32(__rev0_123, -__rev1_123, __rev2_123, __p3_123); \
-  __ret_123 = __builtin_shufflevector(__ret_123, __ret_123, 1, 0); \
-  __ret_123; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
   float64x2_t __ret;
   __ret = vfmaq_f64(__p0, -__p1, (float64x2_t) {__p2, __p2});
@@ -55078,20 +46574,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
   return __ret;
 }
-#else
-__ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
   float32x2_t __ret;
@@ -55144,28 +46631,12 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64_t __ret; \
   __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -55212,28 +46683,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
   poly64x1_t __ret;
@@ -55264,20 +46719,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64(__p0) __extension__ ({ \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64(__p0) __extension__ ({ \
   poly64x2_t __ret; \
@@ -55308,34 +46754,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64(__p0) __extension__ ({ \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_p64(__p0) __extension__ ({ \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
   __ret; \
 })
-#else
-#define vld1_dup_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_dup_p64(__p0) __extension__ ({ \
   poly64x2_t __ret; \
@@ -55366,36 +46794,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_f64(__p0) __extension__ ({ \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
   __ret; \
 })
-#else
-#define vld1_dup_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -55432,36 +46841,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x2(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x2(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x2(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55496,34 +46886,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x2(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x2(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x3(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x3(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x3(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -55560,34 +46932,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x3(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x3(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x4(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x4(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x4(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -55626,34 +46980,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x4(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x4(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_p64(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld2_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_p64(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55722,34 +47058,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_f64(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld2_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_p64(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld2_dup_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_dup_p64(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55784,36 +47102,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_f64(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld2_dup_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
   __ret; \
 })
-#else
-#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x2_t __s1 = __p1; \
@@ -55968,68 +47267,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
   __ret; \
 })
-#else
-#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
   __ret; \
 })
-#else
-#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
   __ret; \
 })
-#else
-#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_p64(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld3_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_p64(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -56102,34 +47362,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_f64(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld3_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_p64(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld3_dup_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_dup_p64(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -56166,36 +47408,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_f64(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld3_dup_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
   __ret; \
 })
-#else
-#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x3_t __s1 = __p1; \
@@ -56364,68 +47587,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
   __ret; \
 })
-#else
-#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
   __ret; \
 })
-#else
-#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
   __ret; \
 })
-#else
-#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_p64(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld4_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_p64(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -56502,34 +47686,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_f64(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld4_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_p64(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld4_dup_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_dup_p64(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -56568,36 +47734,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_f64(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld4_dup_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
   __ret; \
 })
-#else
-#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x4_t __s1 = __p1; \
@@ -56780,68 +47927,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
   __ret; \
 })
-#else
-#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
   __ret; \
 })
-#else
-#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
   __ret; \
 })
-#else
-#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vldrq_p128(__p0) __extension__ ({ \
   poly128_t __ret; \
   __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
   __ret; \
 })
-#else
-#define vldrq_p128(__p0) __extension__ ({ \
-  poly128_t __ret; \
-  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -56859,20 +47967,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
   float64_t __ret;
@@ -57160,20 +48259,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64_t vminnmvq_f64(float64x2_t __p0) {
   float64_t __ret;
@@ -57462,20 +48552,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = __p0 + __p1 * __p2;
   return __ret;
 }
-#else
-__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
   uint32x4_t __s0 = __p0; \
@@ -58039,20 +49120,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = __p0 - __p1 * __p2;
   return __ret;
 }
-#else
-__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
   uint32x4_t __s0 = __p0; \
@@ -58598,20 +49670,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vmov_n_p64(poly64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
   poly64x2_t __ret;
@@ -58642,162 +49705,153 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmov_n_f64(float64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) {__p0};
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_120) {
+  uint16x8_t __ret_120;
+  uint8x8_t __a1_120 = vget_high_u8(__p0_120);
+  __ret_120 = (uint16x8_t)(vshll_n_u8(__a1_120, 0));
+  return __ret_120;
+}
 #else
-__ai float64x1_t vmov_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_121) {
+  uint8x16_t __rev0_121;  __rev0_121 = __builtin_shufflevector(__p0_121, __p0_121, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret_121;
+  uint8x8_t __a1_121 = __noswap_vget_high_u8(__rev0_121);
+  __ret_121 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_121, 0));
+  __ret_121 = __builtin_shufflevector(__ret_121, __ret_121, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret_121;
+}
+__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_122) {
+  uint16x8_t __ret_122;
+  uint8x8_t __a1_122 = __noswap_vget_high_u8(__p0_122);
+  __ret_122 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_122, 0));
+  return __ret_122;
 }
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_124) {
-  uint16x8_t __ret_124;
-  uint8x8_t __a1_124 = vget_high_u8(__p0_124);
-  __ret_124 = (uint16x8_t)(vshll_n_u8(__a1_124, 0));
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_123) {
+  uint64x2_t __ret_123;
+  uint32x2_t __a1_123 = vget_high_u32(__p0_123);
+  __ret_123 = (uint64x2_t)(vshll_n_u32(__a1_123, 0));
+  return __ret_123;
+}
+#else
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_124) {
+  uint32x4_t __rev0_124;  __rev0_124 = __builtin_shufflevector(__p0_124, __p0_124, 3, 2, 1, 0);
+  uint64x2_t __ret_124;
+  uint32x2_t __a1_124 = __noswap_vget_high_u32(__rev0_124);
+  __ret_124 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_124, 0));
+  __ret_124 = __builtin_shufflevector(__ret_124, __ret_124, 1, 0);
   return __ret_124;
 }
-#else
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_125) {
-  uint8x16_t __rev0_125;  __rev0_125 = __builtin_shufflevector(__p0_125, __p0_125, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret_125;
-  uint8x8_t __a1_125 = __noswap_vget_high_u8(__rev0_125);
-  __ret_125 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_125, 0));
-  __ret_125 = __builtin_shufflevector(__ret_125, __ret_125, 7, 6, 5, 4, 3, 2, 1, 0);
+__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_125) {
+  uint64x2_t __ret_125;
+  uint32x2_t __a1_125 = __noswap_vget_high_u32(__p0_125);
+  __ret_125 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_125, 0));
   return __ret_125;
 }
-__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_126) {
-  uint16x8_t __ret_126;
-  uint8x8_t __a1_126 = __noswap_vget_high_u8(__p0_126);
-  __ret_126 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_126, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_126) {
+  uint32x4_t __ret_126;
+  uint16x4_t __a1_126 = vget_high_u16(__p0_126);
+  __ret_126 = (uint32x4_t)(vshll_n_u16(__a1_126, 0));
   return __ret_126;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_127) {
-  uint64x2_t __ret_127;
-  uint32x2_t __a1_127 = vget_high_u32(__p0_127);
-  __ret_127 = (uint64x2_t)(vshll_n_u32(__a1_127, 0));
+#else
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_127) {
+  uint16x8_t __rev0_127;  __rev0_127 = __builtin_shufflevector(__p0_127, __p0_127, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret_127;
+  uint16x4_t __a1_127 = __noswap_vget_high_u16(__rev0_127);
+  __ret_127 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_127, 0));
+  __ret_127 = __builtin_shufflevector(__ret_127, __ret_127, 3, 2, 1, 0);
   return __ret_127;
 }
-#else
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_128) {
-  uint32x4_t __rev0_128;  __rev0_128 = __builtin_shufflevector(__p0_128, __p0_128, 3, 2, 1, 0);
-  uint64x2_t __ret_128;
-  uint32x2_t __a1_128 = __noswap_vget_high_u32(__rev0_128);
-  __ret_128 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_128, 0));
-  __ret_128 = __builtin_shufflevector(__ret_128, __ret_128, 1, 0);
+__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_128) {
+  uint32x4_t __ret_128;
+  uint16x4_t __a1_128 = __noswap_vget_high_u16(__p0_128);
+  __ret_128 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_128, 0));
   return __ret_128;
 }
-__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_129) {
-  uint64x2_t __ret_129;
-  uint32x2_t __a1_129 = __noswap_vget_high_u32(__p0_129);
-  __ret_129 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_129, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_129) {
+  int16x8_t __ret_129;
+  int8x8_t __a1_129 = vget_high_s8(__p0_129);
+  __ret_129 = (int16x8_t)(vshll_n_s8(__a1_129, 0));
   return __ret_129;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_130) {
-  uint32x4_t __ret_130;
-  uint16x4_t __a1_130 = vget_high_u16(__p0_130);
-  __ret_130 = (uint32x4_t)(vshll_n_u16(__a1_130, 0));
+#else
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_130) {
+  int8x16_t __rev0_130;  __rev0_130 = __builtin_shufflevector(__p0_130, __p0_130, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret_130;
+  int8x8_t __a1_130 = __noswap_vget_high_s8(__rev0_130);
+  __ret_130 = (int16x8_t)(__noswap_vshll_n_s8(__a1_130, 0));
+  __ret_130 = __builtin_shufflevector(__ret_130, __ret_130, 7, 6, 5, 4, 3, 2, 1, 0);
   return __ret_130;
 }
-#else
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_131) {
-  uint16x8_t __rev0_131;  __rev0_131 = __builtin_shufflevector(__p0_131, __p0_131, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret_131;
-  uint16x4_t __a1_131 = __noswap_vget_high_u16(__rev0_131);
-  __ret_131 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_131, 0));
-  __ret_131 = __builtin_shufflevector(__ret_131, __ret_131, 3, 2, 1, 0);
+__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_131) {
+  int16x8_t __ret_131;
+  int8x8_t __a1_131 = __noswap_vget_high_s8(__p0_131);
+  __ret_131 = (int16x8_t)(__noswap_vshll_n_s8(__a1_131, 0));
   return __ret_131;
 }
-__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_132) {
-  uint32x4_t __ret_132;
-  uint16x4_t __a1_132 = __noswap_vget_high_u16(__p0_132);
-  __ret_132 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_132, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_132) {
+  int64x2_t __ret_132;
+  int32x2_t __a1_132 = vget_high_s32(__p0_132);
+  __ret_132 = (int64x2_t)(vshll_n_s32(__a1_132, 0));
   return __ret_132;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_133) {
-  int16x8_t __ret_133;
-  int8x8_t __a1_133 = vget_high_s8(__p0_133);
-  __ret_133 = (int16x8_t)(vshll_n_s8(__a1_133, 0));
+#else
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_133) {
+  int32x4_t __rev0_133;  __rev0_133 = __builtin_shufflevector(__p0_133, __p0_133, 3, 2, 1, 0);
+  int64x2_t __ret_133;
+  int32x2_t __a1_133 = __noswap_vget_high_s32(__rev0_133);
+  __ret_133 = (int64x2_t)(__noswap_vshll_n_s32(__a1_133, 0));
+  __ret_133 = __builtin_shufflevector(__ret_133, __ret_133, 1, 0);
   return __ret_133;
 }
-#else
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_134) {
-  int8x16_t __rev0_134;  __rev0_134 = __builtin_shufflevector(__p0_134, __p0_134, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret_134;
-  int8x8_t __a1_134 = __noswap_vget_high_s8(__rev0_134);
-  __ret_134 = (int16x8_t)(__noswap_vshll_n_s8(__a1_134, 0));
-  __ret_134 = __builtin_shufflevector(__ret_134, __ret_134, 7, 6, 5, 4, 3, 2, 1, 0);
+__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_134) {
+  int64x2_t __ret_134;
+  int32x2_t __a1_134 = __noswap_vget_high_s32(__p0_134);
+  __ret_134 = (int64x2_t)(__noswap_vshll_n_s32(__a1_134, 0));
   return __ret_134;
 }
-__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_135) {
-  int16x8_t __ret_135;
-  int8x8_t __a1_135 = __noswap_vget_high_s8(__p0_135);
-  __ret_135 = (int16x8_t)(__noswap_vshll_n_s8(__a1_135, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_135) {
+  int32x4_t __ret_135;
+  int16x4_t __a1_135 = vget_high_s16(__p0_135);
+  __ret_135 = (int32x4_t)(vshll_n_s16(__a1_135, 0));
   return __ret_135;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_136) {
-  int64x2_t __ret_136;
-  int32x2_t __a1_136 = vget_high_s32(__p0_136);
-  __ret_136 = (int64x2_t)(vshll_n_s32(__a1_136, 0));
+#else
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_136) {
+  int16x8_t __rev0_136;  __rev0_136 = __builtin_shufflevector(__p0_136, __p0_136, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret_136;
+  int16x4_t __a1_136 = __noswap_vget_high_s16(__rev0_136);
+  __ret_136 = (int32x4_t)(__noswap_vshll_n_s16(__a1_136, 0));
+  __ret_136 = __builtin_shufflevector(__ret_136, __ret_136, 3, 2, 1, 0);
   return __ret_136;
 }
-#else
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_137) {
-  int32x4_t __rev0_137;  __rev0_137 = __builtin_shufflevector(__p0_137, __p0_137, 3, 2, 1, 0);
-  int64x2_t __ret_137;
-  int32x2_t __a1_137 = __noswap_vget_high_s32(__rev0_137);
-  __ret_137 = (int64x2_t)(__noswap_vshll_n_s32(__a1_137, 0));
-  __ret_137 = __builtin_shufflevector(__ret_137, __ret_137, 1, 0);
+__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_137) {
+  int32x4_t __ret_137;
+  int16x4_t __a1_137 = __noswap_vget_high_s16(__p0_137);
+  __ret_137 = (int32x4_t)(__noswap_vshll_n_s16(__a1_137, 0));
   return __ret_137;
 }
-__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_138) {
-  int64x2_t __ret_138;
-  int32x2_t __a1_138 = __noswap_vget_high_s32(__p0_138);
-  __ret_138 = (int64x2_t)(__noswap_vshll_n_s32(__a1_138, 0));
-  return __ret_138;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_139) {
-  int32x4_t __ret_139;
-  int16x4_t __a1_139 = vget_high_s16(__p0_139);
-  __ret_139 = (int32x4_t)(vshll_n_s16(__a1_139, 0));
-  return __ret_139;
-}
-#else
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_140) {
-  int16x8_t __rev0_140;  __rev0_140 = __builtin_shufflevector(__p0_140, __p0_140, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret_140;
-  int16x4_t __a1_140 = __noswap_vget_high_s16(__rev0_140);
-  __ret_140 = (int32x4_t)(__noswap_vshll_n_s16(__a1_140, 0));
-  __ret_140 = __builtin_shufflevector(__ret_140, __ret_140, 3, 2, 1, 0);
-  return __ret_140;
-}
-__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_141) {
-  int32x4_t __ret_141;
-  int16x4_t __a1_141 = __noswap_vget_high_s16(__p0_141);
-  __ret_141 = (int32x4_t)(__noswap_vshll_n_s16(__a1_141, 0));
-  return __ret_141;
-}
 #endif
 
 #ifdef __LITTLE_ENDIAN__
@@ -58919,58 +49973,37 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 * __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#endif
-
+#define vmuld_lane_f64(__p0_138, __p1_138, __p2_138) __extension__ ({ \
+  float64_t __s0_138 = __p0_138; \
+  float64x1_t __s1_138 = __p1_138; \
+  float64_t __ret_138; \
+  __ret_138 = __s0_138 * vget_lane_f64(__s1_138, __p2_138); \
+  __ret_138; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmuld_lane_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
-  float64_t __s0_142 = __p0_142; \
-  float64x1_t __s1_142 = __p1_142; \
-  float64_t __ret_142; \
-  __ret_142 = __s0_142 * vget_lane_f64(__s1_142, __p2_142); \
-  __ret_142; \
+#define vmuls_lane_f32(__p0_139, __p1_139, __p2_139) __extension__ ({ \
+  float32_t __s0_139 = __p0_139; \
+  float32x2_t __s1_139 = __p1_139; \
+  float32_t __ret_139; \
+  __ret_139 = __s0_139 * vget_lane_f32(__s1_139, __p2_139); \
+  __ret_139; \
 })
 #else
-#define vmuld_lane_f64(__p0_143, __p1_143, __p2_143) __extension__ ({ \
-  float64_t __s0_143 = __p0_143; \
-  float64x1_t __s1_143 = __p1_143; \
-  float64_t __ret_143; \
-  __ret_143 = __s0_143 * __noswap_vget_lane_f64(__s1_143, __p2_143); \
-  __ret_143; \
+#define vmuls_lane_f32(__p0_140, __p1_140, __p2_140) __extension__ ({ \
+  float32_t __s0_140 = __p0_140; \
+  float32x2_t __s1_140 = __p1_140; \
+  float32x2_t __rev1_140;  __rev1_140 = __builtin_shufflevector(__s1_140, __s1_140, 1, 0); \
+  float32_t __ret_140; \
+  __ret_140 = __s0_140 * __noswap_vget_lane_f32(__rev1_140, __p2_140); \
+  __ret_140; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vmuls_lane_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
-  float32_t __s0_144 = __p0_144; \
-  float32x2_t __s1_144 = __p1_144; \
-  float32_t __ret_144; \
-  __ret_144 = __s0_144 * vget_lane_f32(__s1_144, __p2_144); \
-  __ret_144; \
-})
-#else
-#define vmuls_lane_f32(__p0_145, __p1_145, __p2_145) __extension__ ({ \
-  float32_t __s0_145 = __p0_145; \
-  float32x2_t __s1_145 = __p1_145; \
-  float32x2_t __rev1_145;  __rev1_145 = __builtin_shufflevector(__s1_145, __s1_145, 1, 0); \
-  float32_t __ret_145; \
-  __ret_145 = __s0_145 * __noswap_vget_lane_f32(__rev1_145, __p2_145); \
-  __ret_145; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -58978,16 +50011,6 @@
   __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -59009,40 +50032,40 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmuld_laneq_f64(__p0_146, __p1_146, __p2_146) __extension__ ({ \
-  float64_t __s0_146 = __p0_146; \
-  float64x2_t __s1_146 = __p1_146; \
-  float64_t __ret_146; \
-  __ret_146 = __s0_146 * vgetq_lane_f64(__s1_146, __p2_146); \
-  __ret_146; \
+#define vmuld_laneq_f64(__p0_141, __p1_141, __p2_141) __extension__ ({ \
+  float64_t __s0_141 = __p0_141; \
+  float64x2_t __s1_141 = __p1_141; \
+  float64_t __ret_141; \
+  __ret_141 = __s0_141 * vgetq_lane_f64(__s1_141, __p2_141); \
+  __ret_141; \
 })
 #else
-#define vmuld_laneq_f64(__p0_147, __p1_147, __p2_147) __extension__ ({ \
-  float64_t __s0_147 = __p0_147; \
-  float64x2_t __s1_147 = __p1_147; \
-  float64x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
-  float64_t __ret_147; \
-  __ret_147 = __s0_147 * __noswap_vgetq_lane_f64(__rev1_147, __p2_147); \
-  __ret_147; \
+#define vmuld_laneq_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
+  float64_t __s0_142 = __p0_142; \
+  float64x2_t __s1_142 = __p1_142; \
+  float64x2_t __rev1_142;  __rev1_142 = __builtin_shufflevector(__s1_142, __s1_142, 1, 0); \
+  float64_t __ret_142; \
+  __ret_142 = __s0_142 * __noswap_vgetq_lane_f64(__rev1_142, __p2_142); \
+  __ret_142; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmuls_laneq_f32(__p0_148, __p1_148, __p2_148) __extension__ ({ \
-  float32_t __s0_148 = __p0_148; \
-  float32x4_t __s1_148 = __p1_148; \
-  float32_t __ret_148; \
-  __ret_148 = __s0_148 * vgetq_lane_f32(__s1_148, __p2_148); \
-  __ret_148; \
+#define vmuls_laneq_f32(__p0_143, __p1_143, __p2_143) __extension__ ({ \
+  float32_t __s0_143 = __p0_143; \
+  float32x4_t __s1_143 = __p1_143; \
+  float32_t __ret_143; \
+  __ret_143 = __s0_143 * vgetq_lane_f32(__s1_143, __p2_143); \
+  __ret_143; \
 })
 #else
-#define vmuls_laneq_f32(__p0_149, __p1_149, __p2_149) __extension__ ({ \
-  float32_t __s0_149 = __p0_149; \
-  float32x4_t __s1_149 = __p1_149; \
-  float32x4_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 3, 2, 1, 0); \
-  float32_t __ret_149; \
-  __ret_149 = __s0_149 * __noswap_vgetq_lane_f32(__rev1_149, __p2_149); \
-  __ret_149; \
+#define vmuls_laneq_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
+  float32_t __s0_144 = __p0_144; \
+  float32x4_t __s1_144 = __p1_144; \
+  float32x4_t __rev1_144;  __rev1_144 = __builtin_shufflevector(__s1_144, __s1_144, 3, 2, 1, 0); \
+  float32_t __ret_144; \
+  __ret_144 = __s0_144 * __noswap_vgetq_lane_f32(__rev1_144, __p2_144); \
+  __ret_144; \
 })
 #endif
 
@@ -59296,20 +50319,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
   return __ret;
 }
-#else
-__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
   float64x2_t __ret;
@@ -59326,25 +50340,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
   poly128_t __ret;
   __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
   return __ret;
 }
-#else
-__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-__ai poly128_t __noswap_vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
   poly16x8_t __ret;
@@ -59475,7 +50475,7 @@
   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
   poly128_t __ret;
-  __ret = __noswap_vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
+  __ret = vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
   return __ret;
 }
 #endif
@@ -59840,20 +50840,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
   float32x2_t __ret;
@@ -59876,78 +50867,39 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-__ai float64_t __noswap_vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-__ai float32_t __noswap_vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
+#define vmulxd_lane_f64(__p0_145, __p1_145, __p2_145) __extension__ ({ \
+  float64_t __s0_145 = __p0_145; \
+  float64x1_t __s1_145 = __p1_145; \
+  float64_t __ret_145; \
+  __ret_145 = vmulxd_f64(__s0_145, vget_lane_f64(__s1_145, __p2_145)); \
+  __ret_145; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmulxd_lane_f64(__p0_150, __p1_150, __p2_150) __extension__ ({ \
-  float64_t __s0_150 = __p0_150; \
-  float64x1_t __s1_150 = __p1_150; \
-  float64_t __ret_150; \
-  __ret_150 = vmulxd_f64(__s0_150, vget_lane_f64(__s1_150, __p2_150)); \
-  __ret_150; \
+#define vmulxs_lane_f32(__p0_146, __p1_146, __p2_146) __extension__ ({ \
+  float32_t __s0_146 = __p0_146; \
+  float32x2_t __s1_146 = __p1_146; \
+  float32_t __ret_146; \
+  __ret_146 = vmulxs_f32(__s0_146, vget_lane_f32(__s1_146, __p2_146)); \
+  __ret_146; \
 })
 #else
-#define vmulxd_lane_f64(__p0_151, __p1_151, __p2_151) __extension__ ({ \
-  float64_t __s0_151 = __p0_151; \
-  float64x1_t __s1_151 = __p1_151; \
-  float64_t __ret_151; \
-  __ret_151 = __noswap_vmulxd_f64(__s0_151, __noswap_vget_lane_f64(__s1_151, __p2_151)); \
-  __ret_151; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxs_lane_f32(__p0_152, __p1_152, __p2_152) __extension__ ({ \
-  float32_t __s0_152 = __p0_152; \
-  float32x2_t __s1_152 = __p1_152; \
-  float32_t __ret_152; \
-  __ret_152 = vmulxs_f32(__s0_152, vget_lane_f32(__s1_152, __p2_152)); \
-  __ret_152; \
-})
-#else
-#define vmulxs_lane_f32(__p0_153, __p1_153, __p2_153) __extension__ ({ \
-  float32_t __s0_153 = __p0_153; \
-  float32x2_t __s1_153 = __p1_153; \
-  float32x2_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 1, 0); \
-  float32_t __ret_153; \
-  __ret_153 = __noswap_vmulxs_f32(__s0_153, __noswap_vget_lane_f32(__rev1_153, __p2_153)); \
-  __ret_153; \
+#define vmulxs_lane_f32(__p0_147, __p1_147, __p2_147) __extension__ ({ \
+  float32_t __s0_147 = __p0_147; \
+  float32x2_t __s1_147 = __p1_147; \
+  float32x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
+  float32_t __ret_147; \
+  __ret_147 = vmulxs_f32(__s0_147, __noswap_vget_lane_f32(__rev1_147, __p2_147)); \
+  __ret_147; \
 })
 #endif
 
@@ -60014,40 +50966,40 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulxd_laneq_f64(__p0_154, __p1_154, __p2_154) __extension__ ({ \
-  float64_t __s0_154 = __p0_154; \
-  float64x2_t __s1_154 = __p1_154; \
-  float64_t __ret_154; \
-  __ret_154 = vmulxd_f64(__s0_154, vgetq_lane_f64(__s1_154, __p2_154)); \
-  __ret_154; \
+#define vmulxd_laneq_f64(__p0_148, __p1_148, __p2_148) __extension__ ({ \
+  float64_t __s0_148 = __p0_148; \
+  float64x2_t __s1_148 = __p1_148; \
+  float64_t __ret_148; \
+  __ret_148 = vmulxd_f64(__s0_148, vgetq_lane_f64(__s1_148, __p2_148)); \
+  __ret_148; \
 })
 #else
-#define vmulxd_laneq_f64(__p0_155, __p1_155, __p2_155) __extension__ ({ \
-  float64_t __s0_155 = __p0_155; \
-  float64x2_t __s1_155 = __p1_155; \
-  float64x2_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 1, 0); \
-  float64_t __ret_155; \
-  __ret_155 = __noswap_vmulxd_f64(__s0_155, __noswap_vgetq_lane_f64(__rev1_155, __p2_155)); \
-  __ret_155; \
+#define vmulxd_laneq_f64(__p0_149, __p1_149, __p2_149) __extension__ ({ \
+  float64_t __s0_149 = __p0_149; \
+  float64x2_t __s1_149 = __p1_149; \
+  float64x2_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 1, 0); \
+  float64_t __ret_149; \
+  __ret_149 = vmulxd_f64(__s0_149, __noswap_vgetq_lane_f64(__rev1_149, __p2_149)); \
+  __ret_149; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulxs_laneq_f32(__p0_156, __p1_156, __p2_156) __extension__ ({ \
-  float32_t __s0_156 = __p0_156; \
-  float32x4_t __s1_156 = __p1_156; \
-  float32_t __ret_156; \
-  __ret_156 = vmulxs_f32(__s0_156, vgetq_lane_f32(__s1_156, __p2_156)); \
-  __ret_156; \
+#define vmulxs_laneq_f32(__p0_150, __p1_150, __p2_150) __extension__ ({ \
+  float32_t __s0_150 = __p0_150; \
+  float32x4_t __s1_150 = __p1_150; \
+  float32_t __ret_150; \
+  __ret_150 = vmulxs_f32(__s0_150, vgetq_lane_f32(__s1_150, __p2_150)); \
+  __ret_150; \
 })
 #else
-#define vmulxs_laneq_f32(__p0_157, __p1_157, __p2_157) __extension__ ({ \
-  float32_t __s0_157 = __p0_157; \
-  float32x4_t __s1_157 = __p1_157; \
-  float32x4_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 3, 2, 1, 0); \
-  float32_t __ret_157; \
-  __ret_157 = __noswap_vmulxs_f32(__s0_157, __noswap_vgetq_lane_f32(__rev1_157, __p2_157)); \
-  __ret_157; \
+#define vmulxs_laneq_f32(__p0_151, __p1_151, __p2_151) __extension__ ({ \
+  float32_t __s0_151 = __p0_151; \
+  float32x4_t __s1_151 = __p1_151; \
+  float32x4_t __rev1_151;  __rev1_151 = __builtin_shufflevector(__s1_151, __s1_151, 3, 2, 1, 0); \
+  float32_t __ret_151; \
+  __ret_151 = vmulxs_f32(__s0_151, __noswap_vgetq_lane_f32(__rev1_151, __p2_151)); \
+  __ret_151; \
 })
 #endif
 
@@ -60146,48 +51098,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vneg_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = -__p0;
   return __ret;
 }
-#else
-__ai float64x1_t vneg_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vneg_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = -__p0;
   return __ret;
 }
-#else
-__ai int64x1_t vneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vnegd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
   uint8x16_t __ret;
@@ -60928,226 +51853,81 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqabs_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqabsb_s8(int8_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqabsb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqabss_s32(int32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqabss_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqabsd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vqabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqabsh_s16(int16_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqabsh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
   int64x2_t __ret;
@@ -61446,34 +52226,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
   int64x2_t __ret;
@@ -61772,121 +52534,93 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_lane_s32(__p0_152, __p1_152, __p2_152) __extension__ ({ \
+  int32_t __s0_152 = __p0_152; \
+  int32x2_t __s1_152 = __p1_152; \
+  int32_t __ret_152; \
+  __ret_152 = vqdmulhs_s32(__s0_152, vget_lane_s32(__s1_152, __p2_152)); \
+  __ret_152; \
+})
 #else
-__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqdmulhs_lane_s32(__p0_153, __p1_153, __p2_153) __extension__ ({ \
+  int32_t __s0_153 = __p0_153; \
+  int32x2_t __s1_153 = __p1_153; \
+  int32x2_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 1, 0); \
+  int32_t __ret_153; \
+  __ret_153 = vqdmulhs_s32(__s0_153, __noswap_vget_lane_s32(__rev1_153, __p2_153)); \
+  __ret_153; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_lane_s32(__p0_158, __p1_158, __p2_158) __extension__ ({ \
-  int32_t __s0_158 = __p0_158; \
-  int32x2_t __s1_158 = __p1_158; \
-  int32_t __ret_158; \
-  __ret_158 = vqdmulhs_s32(__s0_158, vget_lane_s32(__s1_158, __p2_158)); \
+#define vqdmulhh_lane_s16(__p0_154, __p1_154, __p2_154) __extension__ ({ \
+  int16_t __s0_154 = __p0_154; \
+  int16x4_t __s1_154 = __p1_154; \
+  int16_t __ret_154; \
+  __ret_154 = vqdmulhh_s16(__s0_154, vget_lane_s16(__s1_154, __p2_154)); \
+  __ret_154; \
+})
+#else
+#define vqdmulhh_lane_s16(__p0_155, __p1_155, __p2_155) __extension__ ({ \
+  int16_t __s0_155 = __p0_155; \
+  int16x4_t __s1_155 = __p1_155; \
+  int16x4_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 3, 2, 1, 0); \
+  int16_t __ret_155; \
+  __ret_155 = vqdmulhh_s16(__s0_155, __noswap_vget_lane_s16(__rev1_155, __p2_155)); \
+  __ret_155; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_laneq_s32(__p0_156, __p1_156, __p2_156) __extension__ ({ \
+  int32_t __s0_156 = __p0_156; \
+  int32x4_t __s1_156 = __p1_156; \
+  int32_t __ret_156; \
+  __ret_156 = vqdmulhs_s32(__s0_156, vgetq_lane_s32(__s1_156, __p2_156)); \
+  __ret_156; \
+})
+#else
+#define vqdmulhs_laneq_s32(__p0_157, __p1_157, __p2_157) __extension__ ({ \
+  int32_t __s0_157 = __p0_157; \
+  int32x4_t __s1_157 = __p1_157; \
+  int32x4_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 3, 2, 1, 0); \
+  int32_t __ret_157; \
+  __ret_157 = vqdmulhs_s32(__s0_157, __noswap_vgetq_lane_s32(__rev1_157, __p2_157)); \
+  __ret_157; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhh_laneq_s16(__p0_158, __p1_158, __p2_158) __extension__ ({ \
+  int16_t __s0_158 = __p0_158; \
+  int16x8_t __s1_158 = __p1_158; \
+  int16_t __ret_158; \
+  __ret_158 = vqdmulhh_s16(__s0_158, vgetq_lane_s16(__s1_158, __p2_158)); \
   __ret_158; \
 })
 #else
-#define vqdmulhs_lane_s32(__p0_159, __p1_159, __p2_159) __extension__ ({ \
-  int32_t __s0_159 = __p0_159; \
-  int32x2_t __s1_159 = __p1_159; \
-  int32x2_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 1, 0); \
-  int32_t __ret_159; \
-  __ret_159 = __noswap_vqdmulhs_s32(__s0_159, __noswap_vget_lane_s32(__rev1_159, __p2_159)); \
+#define vqdmulhh_laneq_s16(__p0_159, __p1_159, __p2_159) __extension__ ({ \
+  int16_t __s0_159 = __p0_159; \
+  int16x8_t __s1_159 = __p1_159; \
+  int16x8_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_159; \
+  __ret_159 = vqdmulhh_s16(__s0_159, __noswap_vgetq_lane_s16(__rev1_159, __p2_159)); \
   __ret_159; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_lane_s16(__p0_160, __p1_160, __p2_160) __extension__ ({ \
-  int16_t __s0_160 = __p0_160; \
-  int16x4_t __s1_160 = __p1_160; \
-  int16_t __ret_160; \
-  __ret_160 = vqdmulhh_s16(__s0_160, vget_lane_s16(__s1_160, __p2_160)); \
-  __ret_160; \
-})
-#else
-#define vqdmulhh_lane_s16(__p0_161, __p1_161, __p2_161) __extension__ ({ \
-  int16_t __s0_161 = __p0_161; \
-  int16x4_t __s1_161 = __p1_161; \
-  int16x4_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 3, 2, 1, 0); \
-  int16_t __ret_161; \
-  __ret_161 = __noswap_vqdmulhh_s16(__s0_161, __noswap_vget_lane_s16(__rev1_161, __p2_161)); \
-  __ret_161; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_laneq_s32(__p0_162, __p1_162, __p2_162) __extension__ ({ \
-  int32_t __s0_162 = __p0_162; \
-  int32x4_t __s1_162 = __p1_162; \
-  int32_t __ret_162; \
-  __ret_162 = vqdmulhs_s32(__s0_162, vgetq_lane_s32(__s1_162, __p2_162)); \
-  __ret_162; \
-})
-#else
-#define vqdmulhs_laneq_s32(__p0_163, __p1_163, __p2_163) __extension__ ({ \
-  int32_t __s0_163 = __p0_163; \
-  int32x4_t __s1_163 = __p1_163; \
-  int32x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
-  int32_t __ret_163; \
-  __ret_163 = __noswap_vqdmulhs_s32(__s0_163, __noswap_vgetq_lane_s32(__rev1_163, __p2_163)); \
-  __ret_163; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_laneq_s16(__p0_164, __p1_164, __p2_164) __extension__ ({ \
-  int16_t __s0_164 = __p0_164; \
-  int16x8_t __s1_164 = __p1_164; \
-  int16_t __ret_164; \
-  __ret_164 = vqdmulhh_s16(__s0_164, vgetq_lane_s16(__s1_164, __p2_164)); \
-  __ret_164; \
-})
-#else
-#define vqdmulhh_laneq_s16(__p0_165, __p1_165, __p2_165) __extension__ ({ \
-  int16_t __s0_165 = __p0_165; \
-  int16x8_t __s1_165 = __p1_165; \
-  int16x8_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_165; \
-  __ret_165 = __noswap_vqdmulhh_s16(__s0_165, __noswap_vgetq_lane_s16(__rev1_165, __p2_165)); \
-  __ret_165; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x4_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -61970,44 +52704,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-__ai int64_t __noswap_vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
   int64x2_t __ret;
@@ -62159,82 +52865,82 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulls_lane_s32(__p0_166, __p1_166, __p2_166) __extension__ ({ \
-  int32_t __s0_166 = __p0_166; \
-  int32x2_t __s1_166 = __p1_166; \
-  int64_t __ret_166; \
-  __ret_166 = vqdmulls_s32(__s0_166, vget_lane_s32(__s1_166, __p2_166)); \
+#define vqdmulls_lane_s32(__p0_160, __p1_160, __p2_160) __extension__ ({ \
+  int32_t __s0_160 = __p0_160; \
+  int32x2_t __s1_160 = __p1_160; \
+  int64_t __ret_160; \
+  __ret_160 = vqdmulls_s32(__s0_160, vget_lane_s32(__s1_160, __p2_160)); \
+  __ret_160; \
+})
+#else
+#define vqdmulls_lane_s32(__p0_161, __p1_161, __p2_161) __extension__ ({ \
+  int32_t __s0_161 = __p0_161; \
+  int32x2_t __s1_161 = __p1_161; \
+  int32x2_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 1, 0); \
+  int64_t __ret_161; \
+  __ret_161 = vqdmulls_s32(__s0_161, __noswap_vget_lane_s32(__rev1_161, __p2_161)); \
+  __ret_161; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_lane_s16(__p0_162, __p1_162, __p2_162) __extension__ ({ \
+  int16_t __s0_162 = __p0_162; \
+  int16x4_t __s1_162 = __p1_162; \
+  int32_t __ret_162; \
+  __ret_162 = vqdmullh_s16(__s0_162, vget_lane_s16(__s1_162, __p2_162)); \
+  __ret_162; \
+})
+#else
+#define vqdmullh_lane_s16(__p0_163, __p1_163, __p2_163) __extension__ ({ \
+  int16_t __s0_163 = __p0_163; \
+  int16x4_t __s1_163 = __p1_163; \
+  int16x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
+  int32_t __ret_163; \
+  __ret_163 = vqdmullh_s16(__s0_163, __noswap_vget_lane_s16(__rev1_163, __p2_163)); \
+  __ret_163; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulls_laneq_s32(__p0_164, __p1_164, __p2_164) __extension__ ({ \
+  int32_t __s0_164 = __p0_164; \
+  int32x4_t __s1_164 = __p1_164; \
+  int64_t __ret_164; \
+  __ret_164 = vqdmulls_s32(__s0_164, vgetq_lane_s32(__s1_164, __p2_164)); \
+  __ret_164; \
+})
+#else
+#define vqdmulls_laneq_s32(__p0_165, __p1_165, __p2_165) __extension__ ({ \
+  int32_t __s0_165 = __p0_165; \
+  int32x4_t __s1_165 = __p1_165; \
+  int32x4_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 3, 2, 1, 0); \
+  int64_t __ret_165; \
+  __ret_165 = vqdmulls_s32(__s0_165, __noswap_vgetq_lane_s32(__rev1_165, __p2_165)); \
+  __ret_165; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_laneq_s16(__p0_166, __p1_166, __p2_166) __extension__ ({ \
+  int16_t __s0_166 = __p0_166; \
+  int16x8_t __s1_166 = __p1_166; \
+  int32_t __ret_166; \
+  __ret_166 = vqdmullh_s16(__s0_166, vgetq_lane_s16(__s1_166, __p2_166)); \
   __ret_166; \
 })
 #else
-#define vqdmulls_lane_s32(__p0_167, __p1_167, __p2_167) __extension__ ({ \
-  int32_t __s0_167 = __p0_167; \
-  int32x2_t __s1_167 = __p1_167; \
-  int32x2_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 1, 0); \
-  int64_t __ret_167; \
-  __ret_167 = __noswap_vqdmulls_s32(__s0_167, __noswap_vget_lane_s32(__rev1_167, __p2_167)); \
+#define vqdmullh_laneq_s16(__p0_167, __p1_167, __p2_167) __extension__ ({ \
+  int16_t __s0_167 = __p0_167; \
+  int16x8_t __s1_167 = __p1_167; \
+  int16x8_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32_t __ret_167; \
+  __ret_167 = vqdmullh_s16(__s0_167, __noswap_vgetq_lane_s16(__rev1_167, __p2_167)); \
   __ret_167; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmullh_lane_s16(__p0_168, __p1_168, __p2_168) __extension__ ({ \
-  int16_t __s0_168 = __p0_168; \
-  int16x4_t __s1_168 = __p1_168; \
-  int32_t __ret_168; \
-  __ret_168 = vqdmullh_s16(__s0_168, vget_lane_s16(__s1_168, __p2_168)); \
-  __ret_168; \
-})
-#else
-#define vqdmullh_lane_s16(__p0_169, __p1_169, __p2_169) __extension__ ({ \
-  int16_t __s0_169 = __p0_169; \
-  int16x4_t __s1_169 = __p1_169; \
-  int16x4_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 3, 2, 1, 0); \
-  int32_t __ret_169; \
-  __ret_169 = __noswap_vqdmullh_s16(__s0_169, __noswap_vget_lane_s16(__rev1_169, __p2_169)); \
-  __ret_169; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulls_laneq_s32(__p0_170, __p1_170, __p2_170) __extension__ ({ \
-  int32_t __s0_170 = __p0_170; \
-  int32x4_t __s1_170 = __p1_170; \
-  int64_t __ret_170; \
-  __ret_170 = vqdmulls_s32(__s0_170, vgetq_lane_s32(__s1_170, __p2_170)); \
-  __ret_170; \
-})
-#else
-#define vqdmulls_laneq_s32(__p0_171, __p1_171, __p2_171) __extension__ ({ \
-  int32_t __s0_171 = __p0_171; \
-  int32x4_t __s1_171 = __p1_171; \
-  int32x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
-  int64_t __ret_171; \
-  __ret_171 = __noswap_vqdmulls_s32(__s0_171, __noswap_vgetq_lane_s32(__rev1_171, __p2_171)); \
-  __ret_171; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmullh_laneq_s16(__p0_172, __p1_172, __p2_172) __extension__ ({ \
-  int16_t __s0_172 = __p0_172; \
-  int16x8_t __s1_172 = __p1_172; \
-  int32_t __ret_172; \
-  __ret_172 = vqdmullh_s16(__s0_172, vgetq_lane_s16(__s1_172, __p2_172)); \
-  __ret_172; \
-})
-#else
-#define vqdmullh_laneq_s16(__p0_173, __p1_173, __p2_173) __extension__ ({ \
-  int16_t __s0_173 = __p0_173; \
-  int16x8_t __s1_173 = __p1_173; \
-  int16x8_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32_t __ret_173; \
-  __ret_173 = __noswap_vqdmullh_s16(__s0_173, __noswap_vgetq_lane_s16(__rev1_173, __p2_173)); \
-  __ret_173; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x2_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -62276,90 +52982,36 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqmovns_s32(int32_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqmovns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqmovnd_s64(int64_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqmovnd_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqmovnh_s16(int16_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqmovnh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqmovns_u32(uint32_t __p0) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
   return __ret;
 }
-#else
-__ai uint16_t vqmovns_u32(uint32_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqmovnd_u64(uint64_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vqmovnd_u64(uint64_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqmovnh_u16(uint16_t __p0) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
   return __ret;
 }
-#else
-__ai uint8_t vqmovnh_u16(uint16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
   uint16x8_t __ret;
@@ -62462,48 +53114,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqmovuns_s32(int32_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqmovuns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqmovund_s64(int64_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqmovund_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqmovunh_s16(int16_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqmovunh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
   uint16x8_t __ret;
@@ -62571,191 +53196,118 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqneg_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqnegb_s8(int8_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqnegb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqnegs_s32(int32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqnegs_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqnegd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vqnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqnegh_s16(int16_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqnegh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_lane_s32(__p0_168, __p1_168, __p2_168) __extension__ ({ \
+  int32_t __s0_168 = __p0_168; \
+  int32x2_t __s1_168 = __p1_168; \
+  int32_t __ret_168; \
+  __ret_168 = vqrdmulhs_s32(__s0_168, vget_lane_s32(__s1_168, __p2_168)); \
+  __ret_168; \
+})
 #else
-__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqrdmulhs_lane_s32(__p0_169, __p1_169, __p2_169) __extension__ ({ \
+  int32_t __s0_169 = __p0_169; \
+  int32x2_t __s1_169 = __p1_169; \
+  int32x2_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 1, 0); \
+  int32_t __ret_169; \
+  __ret_169 = vqrdmulhs_s32(__s0_169, __noswap_vget_lane_s32(__rev1_169, __p2_169)); \
+  __ret_169; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_lane_s32(__p0_174, __p1_174, __p2_174) __extension__ ({ \
-  int32_t __s0_174 = __p0_174; \
-  int32x2_t __s1_174 = __p1_174; \
-  int32_t __ret_174; \
-  __ret_174 = vqrdmulhs_s32(__s0_174, vget_lane_s32(__s1_174, __p2_174)); \
+#define vqrdmulhh_lane_s16(__p0_170, __p1_170, __p2_170) __extension__ ({ \
+  int16_t __s0_170 = __p0_170; \
+  int16x4_t __s1_170 = __p1_170; \
+  int16_t __ret_170; \
+  __ret_170 = vqrdmulhh_s16(__s0_170, vget_lane_s16(__s1_170, __p2_170)); \
+  __ret_170; \
+})
+#else
+#define vqrdmulhh_lane_s16(__p0_171, __p1_171, __p2_171) __extension__ ({ \
+  int16_t __s0_171 = __p0_171; \
+  int16x4_t __s1_171 = __p1_171; \
+  int16x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
+  int16_t __ret_171; \
+  __ret_171 = vqrdmulhh_s16(__s0_171, __noswap_vget_lane_s16(__rev1_171, __p2_171)); \
+  __ret_171; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_laneq_s32(__p0_172, __p1_172, __p2_172) __extension__ ({ \
+  int32_t __s0_172 = __p0_172; \
+  int32x4_t __s1_172 = __p1_172; \
+  int32_t __ret_172; \
+  __ret_172 = vqrdmulhs_s32(__s0_172, vgetq_lane_s32(__s1_172, __p2_172)); \
+  __ret_172; \
+})
+#else
+#define vqrdmulhs_laneq_s32(__p0_173, __p1_173, __p2_173) __extension__ ({ \
+  int32_t __s0_173 = __p0_173; \
+  int32x4_t __s1_173 = __p1_173; \
+  int32x4_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 3, 2, 1, 0); \
+  int32_t __ret_173; \
+  __ret_173 = vqrdmulhs_s32(__s0_173, __noswap_vgetq_lane_s32(__rev1_173, __p2_173)); \
+  __ret_173; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhh_laneq_s16(__p0_174, __p1_174, __p2_174) __extension__ ({ \
+  int16_t __s0_174 = __p0_174; \
+  int16x8_t __s1_174 = __p1_174; \
+  int16_t __ret_174; \
+  __ret_174 = vqrdmulhh_s16(__s0_174, vgetq_lane_s16(__s1_174, __p2_174)); \
   __ret_174; \
 })
 #else
-#define vqrdmulhs_lane_s32(__p0_175, __p1_175, __p2_175) __extension__ ({ \
-  int32_t __s0_175 = __p0_175; \
-  int32x2_t __s1_175 = __p1_175; \
-  int32x2_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 1, 0); \
-  int32_t __ret_175; \
-  __ret_175 = __noswap_vqrdmulhs_s32(__s0_175, __noswap_vget_lane_s32(__rev1_175, __p2_175)); \
+#define vqrdmulhh_laneq_s16(__p0_175, __p1_175, __p2_175) __extension__ ({ \
+  int16_t __s0_175 = __p0_175; \
+  int16x8_t __s1_175 = __p1_175; \
+  int16x8_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_175; \
+  __ret_175 = vqrdmulhh_s16(__s0_175, __noswap_vgetq_lane_s16(__rev1_175, __p2_175)); \
   __ret_175; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_lane_s16(__p0_176, __p1_176, __p2_176) __extension__ ({ \
-  int16_t __s0_176 = __p0_176; \
-  int16x4_t __s1_176 = __p1_176; \
-  int16_t __ret_176; \
-  __ret_176 = vqrdmulhh_s16(__s0_176, vget_lane_s16(__s1_176, __p2_176)); \
-  __ret_176; \
-})
-#else
-#define vqrdmulhh_lane_s16(__p0_177, __p1_177, __p2_177) __extension__ ({ \
-  int16_t __s0_177 = __p0_177; \
-  int16x4_t __s1_177 = __p1_177; \
-  int16x4_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 3, 2, 1, 0); \
-  int16_t __ret_177; \
-  __ret_177 = __noswap_vqrdmulhh_s16(__s0_177, __noswap_vget_lane_s16(__rev1_177, __p2_177)); \
-  __ret_177; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_laneq_s32(__p0_178, __p1_178, __p2_178) __extension__ ({ \
-  int32_t __s0_178 = __p0_178; \
-  int32x4_t __s1_178 = __p1_178; \
-  int32_t __ret_178; \
-  __ret_178 = vqrdmulhs_s32(__s0_178, vgetq_lane_s32(__s1_178, __p2_178)); \
-  __ret_178; \
-})
-#else
-#define vqrdmulhs_laneq_s32(__p0_179, __p1_179, __p2_179) __extension__ ({ \
-  int32_t __s0_179 = __p0_179; \
-  int32x4_t __s1_179 = __p1_179; \
-  int32x4_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 3, 2, 1, 0); \
-  int32_t __ret_179; \
-  __ret_179 = __noswap_vqrdmulhs_s32(__s0_179, __noswap_vgetq_lane_s32(__rev1_179, __p2_179)); \
-  __ret_179; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_laneq_s16(__p0_180, __p1_180, __p2_180) __extension__ ({ \
-  int16_t __s0_180 = __p0_180; \
-  int16x8_t __s1_180 = __p1_180; \
-  int16_t __ret_180; \
-  __ret_180 = vqrdmulhh_s16(__s0_180, vgetq_lane_s16(__s1_180, __p2_180)); \
-  __ret_180; \
-})
-#else
-#define vqrdmulhh_laneq_s16(__p0_181, __p1_181, __p2_181) __extension__ ({ \
-  int16_t __s0_181 = __p0_181; \
-  int16x8_t __s1_181 = __p1_181; \
-  int16x8_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_181; \
-  __ret_181 = __noswap_vqrdmulhh_s16(__s0_181, __noswap_vgetq_lane_s16(__rev1_181, __p2_181)); \
-  __ret_181; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x4_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -62839,1210 +53391,684 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u32(__p0_176, __p1_176, __p2_176) __extension__ ({ \
+  uint16x4_t __s0_176 = __p0_176; \
+  uint32x4_t __s1_176 = __p1_176; \
+  uint16x8_t __ret_176; \
+  __ret_176 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_176), (uint16x4_t)(vqrshrn_n_u32(__s1_176, __p2_176)))); \
+  __ret_176; \
+})
 #else
-__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqrshrn_high_n_u32(__p0_177, __p1_177, __p2_177) __extension__ ({ \
+  uint16x4_t __s0_177 = __p0_177; \
+  uint32x4_t __s1_177 = __p1_177; \
+  uint16x4_t __rev0_177;  __rev0_177 = __builtin_shufflevector(__s0_177, __s0_177, 3, 2, 1, 0); \
+  uint32x4_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 3, 2, 1, 0); \
+  uint16x8_t __ret_177; \
+  __ret_177 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_177), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_177, __p2_177)))); \
+  __ret_177 = __builtin_shufflevector(__ret_177, __ret_177, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_177; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u32(__p0_182, __p1_182, __p2_182) __extension__ ({ \
-  uint16x4_t __s0_182 = __p0_182; \
-  uint32x4_t __s1_182 = __p1_182; \
-  uint16x8_t __ret_182; \
-  __ret_182 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_182), (uint16x4_t)(vqrshrn_n_u32(__s1_182, __p2_182)))); \
+#define vqrshrn_high_n_u64(__p0_178, __p1_178, __p2_178) __extension__ ({ \
+  uint32x2_t __s0_178 = __p0_178; \
+  uint64x2_t __s1_178 = __p1_178; \
+  uint32x4_t __ret_178; \
+  __ret_178 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_178), (uint32x2_t)(vqrshrn_n_u64(__s1_178, __p2_178)))); \
+  __ret_178; \
+})
+#else
+#define vqrshrn_high_n_u64(__p0_179, __p1_179, __p2_179) __extension__ ({ \
+  uint32x2_t __s0_179 = __p0_179; \
+  uint64x2_t __s1_179 = __p1_179; \
+  uint32x2_t __rev0_179;  __rev0_179 = __builtin_shufflevector(__s0_179, __s0_179, 1, 0); \
+  uint64x2_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 1, 0); \
+  uint32x4_t __ret_179; \
+  __ret_179 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_179), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_179, __p2_179)))); \
+  __ret_179 = __builtin_shufflevector(__ret_179, __ret_179, 3, 2, 1, 0); \
+  __ret_179; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u16(__p0_180, __p1_180, __p2_180) __extension__ ({ \
+  uint8x8_t __s0_180 = __p0_180; \
+  uint16x8_t __s1_180 = __p1_180; \
+  uint8x16_t __ret_180; \
+  __ret_180 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_180), (uint8x8_t)(vqrshrn_n_u16(__s1_180, __p2_180)))); \
+  __ret_180; \
+})
+#else
+#define vqrshrn_high_n_u16(__p0_181, __p1_181, __p2_181) __extension__ ({ \
+  uint8x8_t __s0_181 = __p0_181; \
+  uint16x8_t __s1_181 = __p1_181; \
+  uint8x8_t __rev0_181;  __rev0_181 = __builtin_shufflevector(__s0_181, __s0_181, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_181; \
+  __ret_181 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_181), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_181, __p2_181)))); \
+  __ret_181 = __builtin_shufflevector(__ret_181, __ret_181, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_181; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_s32(__p0_182, __p1_182, __p2_182) __extension__ ({ \
+  int16x4_t __s0_182 = __p0_182; \
+  int32x4_t __s1_182 = __p1_182; \
+  int16x8_t __ret_182; \
+  __ret_182 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_182), (int16x4_t)(vqrshrn_n_s32(__s1_182, __p2_182)))); \
   __ret_182; \
 })
 #else
-#define vqrshrn_high_n_u32(__p0_183, __p1_183, __p2_183) __extension__ ({ \
-  uint16x4_t __s0_183 = __p0_183; \
-  uint32x4_t __s1_183 = __p1_183; \
-  uint16x4_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 3, 2, 1, 0); \
-  uint32x4_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 3, 2, 1, 0); \
-  uint16x8_t __ret_183; \
-  __ret_183 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_183), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_183, __p2_183)))); \
+#define vqrshrn_high_n_s32(__p0_183, __p1_183, __p2_183) __extension__ ({ \
+  int16x4_t __s0_183 = __p0_183; \
+  int32x4_t __s1_183 = __p1_183; \
+  int16x4_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 3, 2, 1, 0); \
+  int32x4_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 3, 2, 1, 0); \
+  int16x8_t __ret_183; \
+  __ret_183 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_183), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_183, __p2_183)))); \
   __ret_183 = __builtin_shufflevector(__ret_183, __ret_183, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_183; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u64(__p0_184, __p1_184, __p2_184) __extension__ ({ \
-  uint32x2_t __s0_184 = __p0_184; \
-  uint64x2_t __s1_184 = __p1_184; \
-  uint32x4_t __ret_184; \
-  __ret_184 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_184), (uint32x2_t)(vqrshrn_n_u64(__s1_184, __p2_184)))); \
+#define vqrshrn_high_n_s64(__p0_184, __p1_184, __p2_184) __extension__ ({ \
+  int32x2_t __s0_184 = __p0_184; \
+  int64x2_t __s1_184 = __p1_184; \
+  int32x4_t __ret_184; \
+  __ret_184 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_184), (int32x2_t)(vqrshrn_n_s64(__s1_184, __p2_184)))); \
   __ret_184; \
 })
 #else
-#define vqrshrn_high_n_u64(__p0_185, __p1_185, __p2_185) __extension__ ({ \
-  uint32x2_t __s0_185 = __p0_185; \
-  uint64x2_t __s1_185 = __p1_185; \
-  uint32x2_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 1, 0); \
-  uint64x2_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 1, 0); \
-  uint32x4_t __ret_185; \
-  __ret_185 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_185), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_185, __p2_185)))); \
+#define vqrshrn_high_n_s64(__p0_185, __p1_185, __p2_185) __extension__ ({ \
+  int32x2_t __s0_185 = __p0_185; \
+  int64x2_t __s1_185 = __p1_185; \
+  int32x2_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 1, 0); \
+  int64x2_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 1, 0); \
+  int32x4_t __ret_185; \
+  __ret_185 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_185), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_185, __p2_185)))); \
   __ret_185 = __builtin_shufflevector(__ret_185, __ret_185, 3, 2, 1, 0); \
   __ret_185; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u16(__p0_186, __p1_186, __p2_186) __extension__ ({ \
-  uint8x8_t __s0_186 = __p0_186; \
-  uint16x8_t __s1_186 = __p1_186; \
-  uint8x16_t __ret_186; \
-  __ret_186 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_186), (uint8x8_t)(vqrshrn_n_u16(__s1_186, __p2_186)))); \
+#define vqrshrn_high_n_s16(__p0_186, __p1_186, __p2_186) __extension__ ({ \
+  int8x8_t __s0_186 = __p0_186; \
+  int16x8_t __s1_186 = __p1_186; \
+  int8x16_t __ret_186; \
+  __ret_186 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_186), (int8x8_t)(vqrshrn_n_s16(__s1_186, __p2_186)))); \
   __ret_186; \
 })
 #else
-#define vqrshrn_high_n_u16(__p0_187, __p1_187, __p2_187) __extension__ ({ \
-  uint8x8_t __s0_187 = __p0_187; \
-  uint16x8_t __s1_187 = __p1_187; \
-  uint8x8_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_187; \
-  __ret_187 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_187), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_187, __p2_187)))); \
+#define vqrshrn_high_n_s16(__p0_187, __p1_187, __p2_187) __extension__ ({ \
+  int8x8_t __s0_187 = __p0_187; \
+  int16x8_t __s1_187 = __p1_187; \
+  int8x8_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_187; \
+  __ret_187 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_187), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_187, __p2_187)))); \
   __ret_187 = __builtin_shufflevector(__ret_187, __ret_187, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_187; \
 })
 #endif
 
+#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s32(__p0_188, __p1_188, __p2_188) __extension__ ({ \
+#define vqrshrun_high_n_s32(__p0_188, __p1_188, __p2_188) __extension__ ({ \
   int16x4_t __s0_188 = __p0_188; \
   int32x4_t __s1_188 = __p1_188; \
   int16x8_t __ret_188; \
-  __ret_188 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_188), (int16x4_t)(vqrshrn_n_s32(__s1_188, __p2_188)))); \
+  __ret_188 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_188), (int16x4_t)(vqrshrun_n_s32(__s1_188, __p2_188)))); \
   __ret_188; \
 })
 #else
-#define vqrshrn_high_n_s32(__p0_189, __p1_189, __p2_189) __extension__ ({ \
+#define vqrshrun_high_n_s32(__p0_189, __p1_189, __p2_189) __extension__ ({ \
   int16x4_t __s0_189 = __p0_189; \
   int32x4_t __s1_189 = __p1_189; \
   int16x4_t __rev0_189;  __rev0_189 = __builtin_shufflevector(__s0_189, __s0_189, 3, 2, 1, 0); \
   int32x4_t __rev1_189;  __rev1_189 = __builtin_shufflevector(__s1_189, __s1_189, 3, 2, 1, 0); \
   int16x8_t __ret_189; \
-  __ret_189 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_189), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_189, __p2_189)))); \
+  __ret_189 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_189), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_189, __p2_189)))); \
   __ret_189 = __builtin_shufflevector(__ret_189, __ret_189, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_189; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s64(__p0_190, __p1_190, __p2_190) __extension__ ({ \
+#define vqrshrun_high_n_s64(__p0_190, __p1_190, __p2_190) __extension__ ({ \
   int32x2_t __s0_190 = __p0_190; \
   int64x2_t __s1_190 = __p1_190; \
   int32x4_t __ret_190; \
-  __ret_190 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_190), (int32x2_t)(vqrshrn_n_s64(__s1_190, __p2_190)))); \
+  __ret_190 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_190), (int32x2_t)(vqrshrun_n_s64(__s1_190, __p2_190)))); \
   __ret_190; \
 })
 #else
-#define vqrshrn_high_n_s64(__p0_191, __p1_191, __p2_191) __extension__ ({ \
+#define vqrshrun_high_n_s64(__p0_191, __p1_191, __p2_191) __extension__ ({ \
   int32x2_t __s0_191 = __p0_191; \
   int64x2_t __s1_191 = __p1_191; \
   int32x2_t __rev0_191;  __rev0_191 = __builtin_shufflevector(__s0_191, __s0_191, 1, 0); \
   int64x2_t __rev1_191;  __rev1_191 = __builtin_shufflevector(__s1_191, __s1_191, 1, 0); \
   int32x4_t __ret_191; \
-  __ret_191 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_191), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_191, __p2_191)))); \
+  __ret_191 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_191), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_191, __p2_191)))); \
   __ret_191 = __builtin_shufflevector(__ret_191, __ret_191, 3, 2, 1, 0); \
   __ret_191; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s16(__p0_192, __p1_192, __p2_192) __extension__ ({ \
+#define vqrshrun_high_n_s16(__p0_192, __p1_192, __p2_192) __extension__ ({ \
   int8x8_t __s0_192 = __p0_192; \
   int16x8_t __s1_192 = __p1_192; \
   int8x16_t __ret_192; \
-  __ret_192 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_192), (int8x8_t)(vqrshrn_n_s16(__s1_192, __p2_192)))); \
+  __ret_192 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_192), (int8x8_t)(vqrshrun_n_s16(__s1_192, __p2_192)))); \
   __ret_192; \
 })
 #else
-#define vqrshrn_high_n_s16(__p0_193, __p1_193, __p2_193) __extension__ ({ \
+#define vqrshrun_high_n_s16(__p0_193, __p1_193, __p2_193) __extension__ ({ \
   int8x8_t __s0_193 = __p0_193; \
   int16x8_t __s1_193 = __p1_193; \
   int8x8_t __rev0_193;  __rev0_193 = __builtin_shufflevector(__s0_193, __s0_193, 7, 6, 5, 4, 3, 2, 1, 0); \
   int16x8_t __rev1_193;  __rev1_193 = __builtin_shufflevector(__s1_193, __s1_193, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x16_t __ret_193; \
-  __ret_193 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_193), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_193, __p2_193)))); \
+  __ret_193 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_193), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_193, __p2_193)))); \
   __ret_193 = __builtin_shufflevector(__ret_193, __ret_193, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_193; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
+  return __ret;
+}
+__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
+  return __ret;
+}
+__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
+  return __ret;
+}
+__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
+  return __ret;
+}
+__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
+  return __ret;
+}
+__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
+  return __ret;
+}
+__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
+  return __ret;
+}
+__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
+  return __ret;
+}
+#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
+  __ret; \
+})
+#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
+  __ret; \
+})
+#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
   int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
   __ret; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
+#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
   int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
   int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
   __ret; \
 })
-#endif
-
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s32(__p0_194, __p1_194, __p2_194) __extension__ ({ \
-  int16x4_t __s0_194 = __p0_194; \
-  int32x4_t __s1_194 = __p1_194; \
-  int16x8_t __ret_194; \
-  __ret_194 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_194), (int16x4_t)(vqrshrun_n_s32(__s1_194, __p2_194)))); \
+#define vqshrn_high_n_u32(__p0_194, __p1_194, __p2_194) __extension__ ({ \
+  uint16x4_t __s0_194 = __p0_194; \
+  uint32x4_t __s1_194 = __p1_194; \
+  uint16x8_t __ret_194; \
+  __ret_194 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_194), (uint16x4_t)(vqshrn_n_u32(__s1_194, __p2_194)))); \
   __ret_194; \
 })
 #else
-#define vqrshrun_high_n_s32(__p0_195, __p1_195, __p2_195) __extension__ ({ \
-  int16x4_t __s0_195 = __p0_195; \
-  int32x4_t __s1_195 = __p1_195; \
-  int16x4_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 3, 2, 1, 0); \
-  int32x4_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 3, 2, 1, 0); \
-  int16x8_t __ret_195; \
-  __ret_195 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_195), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_195, __p2_195)))); \
+#define vqshrn_high_n_u32(__p0_195, __p1_195, __p2_195) __extension__ ({ \
+  uint16x4_t __s0_195 = __p0_195; \
+  uint32x4_t __s1_195 = __p1_195; \
+  uint16x4_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 3, 2, 1, 0); \
+  uint32x4_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 3, 2, 1, 0); \
+  uint16x8_t __ret_195; \
+  __ret_195 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_195), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_195, __p2_195)))); \
   __ret_195 = __builtin_shufflevector(__ret_195, __ret_195, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_195; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s64(__p0_196, __p1_196, __p2_196) __extension__ ({ \
-  int32x2_t __s0_196 = __p0_196; \
-  int64x2_t __s1_196 = __p1_196; \
-  int32x4_t __ret_196; \
-  __ret_196 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_196), (int32x2_t)(vqrshrun_n_s64(__s1_196, __p2_196)))); \
+#define vqshrn_high_n_u64(__p0_196, __p1_196, __p2_196) __extension__ ({ \
+  uint32x2_t __s0_196 = __p0_196; \
+  uint64x2_t __s1_196 = __p1_196; \
+  uint32x4_t __ret_196; \
+  __ret_196 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_196), (uint32x2_t)(vqshrn_n_u64(__s1_196, __p2_196)))); \
   __ret_196; \
 })
 #else
-#define vqrshrun_high_n_s64(__p0_197, __p1_197, __p2_197) __extension__ ({ \
-  int32x2_t __s0_197 = __p0_197; \
-  int64x2_t __s1_197 = __p1_197; \
-  int32x2_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 1, 0); \
-  int64x2_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 1, 0); \
-  int32x4_t __ret_197; \
-  __ret_197 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_197), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_197, __p2_197)))); \
+#define vqshrn_high_n_u64(__p0_197, __p1_197, __p2_197) __extension__ ({ \
+  uint32x2_t __s0_197 = __p0_197; \
+  uint64x2_t __s1_197 = __p1_197; \
+  uint32x2_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 1, 0); \
+  uint64x2_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 1, 0); \
+  uint32x4_t __ret_197; \
+  __ret_197 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_197), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_197, __p2_197)))); \
   __ret_197 = __builtin_shufflevector(__ret_197, __ret_197, 3, 2, 1, 0); \
   __ret_197; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s16(__p0_198, __p1_198, __p2_198) __extension__ ({ \
-  int8x8_t __s0_198 = __p0_198; \
-  int16x8_t __s1_198 = __p1_198; \
-  int8x16_t __ret_198; \
-  __ret_198 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_198), (int8x8_t)(vqrshrun_n_s16(__s1_198, __p2_198)))); \
+#define vqshrn_high_n_u16(__p0_198, __p1_198, __p2_198) __extension__ ({ \
+  uint8x8_t __s0_198 = __p0_198; \
+  uint16x8_t __s1_198 = __p1_198; \
+  uint8x16_t __ret_198; \
+  __ret_198 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_198), (uint8x8_t)(vqshrn_n_u16(__s1_198, __p2_198)))); \
   __ret_198; \
 })
 #else
-#define vqrshrun_high_n_s16(__p0_199, __p1_199, __p2_199) __extension__ ({ \
-  int8x8_t __s0_199 = __p0_199; \
-  int16x8_t __s1_199 = __p1_199; \
-  int8x8_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_199; \
-  __ret_199 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_199), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_199, __p2_199)))); \
+#define vqshrn_high_n_u16(__p0_199, __p1_199, __p2_199) __extension__ ({ \
+  uint8x8_t __s0_199 = __p0_199; \
+  uint16x8_t __s1_199 = __p1_199; \
+  uint8x8_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_199; \
+  __ret_199 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_199), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_199, __p2_199)))); \
   __ret_199 = __builtin_shufflevector(__ret_199, __ret_199, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_199; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u32(__p0_200, __p1_200, __p2_200) __extension__ ({ \
-  uint16x4_t __s0_200 = __p0_200; \
-  uint32x4_t __s1_200 = __p1_200; \
-  uint16x8_t __ret_200; \
-  __ret_200 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_200), (uint16x4_t)(vqshrn_n_u32(__s1_200, __p2_200)))); \
+#define vqshrn_high_n_s32(__p0_200, __p1_200, __p2_200) __extension__ ({ \
+  int16x4_t __s0_200 = __p0_200; \
+  int32x4_t __s1_200 = __p1_200; \
+  int16x8_t __ret_200; \
+  __ret_200 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_200), (int16x4_t)(vqshrn_n_s32(__s1_200, __p2_200)))); \
   __ret_200; \
 })
 #else
-#define vqshrn_high_n_u32(__p0_201, __p1_201, __p2_201) __extension__ ({ \
-  uint16x4_t __s0_201 = __p0_201; \
-  uint32x4_t __s1_201 = __p1_201; \
-  uint16x4_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 3, 2, 1, 0); \
-  uint32x4_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 3, 2, 1, 0); \
-  uint16x8_t __ret_201; \
-  __ret_201 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_201), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_201, __p2_201)))); \
+#define vqshrn_high_n_s32(__p0_201, __p1_201, __p2_201) __extension__ ({ \
+  int16x4_t __s0_201 = __p0_201; \
+  int32x4_t __s1_201 = __p1_201; \
+  int16x4_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 3, 2, 1, 0); \
+  int32x4_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 3, 2, 1, 0); \
+  int16x8_t __ret_201; \
+  __ret_201 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_201), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_201, __p2_201)))); \
   __ret_201 = __builtin_shufflevector(__ret_201, __ret_201, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_201; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u64(__p0_202, __p1_202, __p2_202) __extension__ ({ \
-  uint32x2_t __s0_202 = __p0_202; \
-  uint64x2_t __s1_202 = __p1_202; \
-  uint32x4_t __ret_202; \
-  __ret_202 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_202), (uint32x2_t)(vqshrn_n_u64(__s1_202, __p2_202)))); \
+#define vqshrn_high_n_s64(__p0_202, __p1_202, __p2_202) __extension__ ({ \
+  int32x2_t __s0_202 = __p0_202; \
+  int64x2_t __s1_202 = __p1_202; \
+  int32x4_t __ret_202; \
+  __ret_202 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_202), (int32x2_t)(vqshrn_n_s64(__s1_202, __p2_202)))); \
   __ret_202; \
 })
 #else
-#define vqshrn_high_n_u64(__p0_203, __p1_203, __p2_203) __extension__ ({ \
-  uint32x2_t __s0_203 = __p0_203; \
-  uint64x2_t __s1_203 = __p1_203; \
-  uint32x2_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 1, 0); \
-  uint64x2_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 1, 0); \
-  uint32x4_t __ret_203; \
-  __ret_203 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_203), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_203, __p2_203)))); \
+#define vqshrn_high_n_s64(__p0_203, __p1_203, __p2_203) __extension__ ({ \
+  int32x2_t __s0_203 = __p0_203; \
+  int64x2_t __s1_203 = __p1_203; \
+  int32x2_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 1, 0); \
+  int64x2_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 1, 0); \
+  int32x4_t __ret_203; \
+  __ret_203 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_203), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_203, __p2_203)))); \
   __ret_203 = __builtin_shufflevector(__ret_203, __ret_203, 3, 2, 1, 0); \
   __ret_203; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u16(__p0_204, __p1_204, __p2_204) __extension__ ({ \
-  uint8x8_t __s0_204 = __p0_204; \
-  uint16x8_t __s1_204 = __p1_204; \
-  uint8x16_t __ret_204; \
-  __ret_204 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_204), (uint8x8_t)(vqshrn_n_u16(__s1_204, __p2_204)))); \
+#define vqshrn_high_n_s16(__p0_204, __p1_204, __p2_204) __extension__ ({ \
+  int8x8_t __s0_204 = __p0_204; \
+  int16x8_t __s1_204 = __p1_204; \
+  int8x16_t __ret_204; \
+  __ret_204 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_204), (int8x8_t)(vqshrn_n_s16(__s1_204, __p2_204)))); \
   __ret_204; \
 })
 #else
-#define vqshrn_high_n_u16(__p0_205, __p1_205, __p2_205) __extension__ ({ \
-  uint8x8_t __s0_205 = __p0_205; \
-  uint16x8_t __s1_205 = __p1_205; \
-  uint8x8_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_205; \
-  __ret_205 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_205), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_205, __p2_205)))); \
+#define vqshrn_high_n_s16(__p0_205, __p1_205, __p2_205) __extension__ ({ \
+  int8x8_t __s0_205 = __p0_205; \
+  int16x8_t __s1_205 = __p1_205; \
+  int8x8_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_205; \
+  __ret_205 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_205), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_205, __p2_205)))); \
   __ret_205 = __builtin_shufflevector(__ret_205, __ret_205, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_205; \
 })
 #endif
 
+#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s32(__p0_206, __p1_206, __p2_206) __extension__ ({ \
+#define vqshrun_high_n_s32(__p0_206, __p1_206, __p2_206) __extension__ ({ \
   int16x4_t __s0_206 = __p0_206; \
   int32x4_t __s1_206 = __p1_206; \
   int16x8_t __ret_206; \
-  __ret_206 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_206), (int16x4_t)(vqshrn_n_s32(__s1_206, __p2_206)))); \
+  __ret_206 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_206), (int16x4_t)(vqshrun_n_s32(__s1_206, __p2_206)))); \
   __ret_206; \
 })
 #else
-#define vqshrn_high_n_s32(__p0_207, __p1_207, __p2_207) __extension__ ({ \
+#define vqshrun_high_n_s32(__p0_207, __p1_207, __p2_207) __extension__ ({ \
   int16x4_t __s0_207 = __p0_207; \
   int32x4_t __s1_207 = __p1_207; \
   int16x4_t __rev0_207;  __rev0_207 = __builtin_shufflevector(__s0_207, __s0_207, 3, 2, 1, 0); \
   int32x4_t __rev1_207;  __rev1_207 = __builtin_shufflevector(__s1_207, __s1_207, 3, 2, 1, 0); \
   int16x8_t __ret_207; \
-  __ret_207 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_207), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_207, __p2_207)))); \
+  __ret_207 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_207), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_207, __p2_207)))); \
   __ret_207 = __builtin_shufflevector(__ret_207, __ret_207, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_207; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s64(__p0_208, __p1_208, __p2_208) __extension__ ({ \
+#define vqshrun_high_n_s64(__p0_208, __p1_208, __p2_208) __extension__ ({ \
   int32x2_t __s0_208 = __p0_208; \
   int64x2_t __s1_208 = __p1_208; \
   int32x4_t __ret_208; \
-  __ret_208 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_208), (int32x2_t)(vqshrn_n_s64(__s1_208, __p2_208)))); \
+  __ret_208 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_208), (int32x2_t)(vqshrun_n_s64(__s1_208, __p2_208)))); \
   __ret_208; \
 })
 #else
-#define vqshrn_high_n_s64(__p0_209, __p1_209, __p2_209) __extension__ ({ \
+#define vqshrun_high_n_s64(__p0_209, __p1_209, __p2_209) __extension__ ({ \
   int32x2_t __s0_209 = __p0_209; \
   int64x2_t __s1_209 = __p1_209; \
   int32x2_t __rev0_209;  __rev0_209 = __builtin_shufflevector(__s0_209, __s0_209, 1, 0); \
   int64x2_t __rev1_209;  __rev1_209 = __builtin_shufflevector(__s1_209, __s1_209, 1, 0); \
   int32x4_t __ret_209; \
-  __ret_209 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_209), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_209, __p2_209)))); \
+  __ret_209 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_209), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_209, __p2_209)))); \
   __ret_209 = __builtin_shufflevector(__ret_209, __ret_209, 3, 2, 1, 0); \
   __ret_209; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s16(__p0_210, __p1_210, __p2_210) __extension__ ({ \
+#define vqshrun_high_n_s16(__p0_210, __p1_210, __p2_210) __extension__ ({ \
   int8x8_t __s0_210 = __p0_210; \
   int16x8_t __s1_210 = __p1_210; \
   int8x16_t __ret_210; \
-  __ret_210 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_210), (int8x8_t)(vqshrn_n_s16(__s1_210, __p2_210)))); \
+  __ret_210 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_210), (int8x8_t)(vqshrun_n_s16(__s1_210, __p2_210)))); \
   __ret_210; \
 })
 #else
-#define vqshrn_high_n_s16(__p0_211, __p1_211, __p2_211) __extension__ ({ \
+#define vqshrun_high_n_s16(__p0_211, __p1_211, __p2_211) __extension__ ({ \
   int8x8_t __s0_211 = __p0_211; \
   int16x8_t __s1_211 = __p1_211; \
   int8x8_t __rev0_211;  __rev0_211 = __builtin_shufflevector(__s0_211, __s0_211, 7, 6, 5, 4, 3, 2, 1, 0); \
   int16x8_t __rev1_211;  __rev1_211 = __builtin_shufflevector(__s1_211, __s1_211, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x16_t __ret_211; \
-  __ret_211 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_211), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_211, __p2_211)))); \
+  __ret_211 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_211), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_211, __p2_211)))); \
   __ret_211 = __builtin_shufflevector(__ret_211, __ret_211, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_211; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s32(__p0_212, __p1_212, __p2_212) __extension__ ({ \
-  int16x4_t __s0_212 = __p0_212; \
-  int32x4_t __s1_212 = __p1_212; \
-  int16x8_t __ret_212; \
-  __ret_212 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_212), (int16x4_t)(vqshrun_n_s32(__s1_212, __p2_212)))); \
-  __ret_212; \
-})
-#else
-#define vqshrun_high_n_s32(__p0_213, __p1_213, __p2_213) __extension__ ({ \
-  int16x4_t __s0_213 = __p0_213; \
-  int32x4_t __s1_213 = __p1_213; \
-  int16x4_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 3, 2, 1, 0); \
-  int32x4_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 3, 2, 1, 0); \
-  int16x8_t __ret_213; \
-  __ret_213 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_213), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_213, __p2_213)))); \
-  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_213; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s64(__p0_214, __p1_214, __p2_214) __extension__ ({ \
-  int32x2_t __s0_214 = __p0_214; \
-  int64x2_t __s1_214 = __p1_214; \
-  int32x4_t __ret_214; \
-  __ret_214 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_214), (int32x2_t)(vqshrun_n_s64(__s1_214, __p2_214)))); \
-  __ret_214; \
-})
-#else
-#define vqshrun_high_n_s64(__p0_215, __p1_215, __p2_215) __extension__ ({ \
-  int32x2_t __s0_215 = __p0_215; \
-  int64x2_t __s1_215 = __p1_215; \
-  int32x2_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 1, 0); \
-  int64x2_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 1, 0); \
-  int32x4_t __ret_215; \
-  __ret_215 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_215), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_215, __p2_215)))); \
-  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 3, 2, 1, 0); \
-  __ret_215; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s16(__p0_216, __p1_216, __p2_216) __extension__ ({ \
-  int8x8_t __s0_216 = __p0_216; \
-  int16x8_t __s1_216 = __p1_216; \
-  int8x16_t __ret_216; \
-  __ret_216 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_216), (int8x8_t)(vqshrun_n_s16(__s1_216, __p2_216)))); \
-  __ret_216; \
-})
-#else
-#define vqshrun_high_n_s16(__p0_217, __p1_217, __p2_217) __extension__ ({ \
-  int8x8_t __s0_217 = __p0_217; \
-  int16x8_t __s1_217 = __p1_217; \
-  int8x8_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_217; \
-  __ret_217 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_217), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_217, __p2_217)))); \
-  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_217; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   int16_t __ret; \
   __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int32_t __ret; \
   __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
   int16_t __s0 = __p0; \
   int8_t __ret; \
   __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
   poly8x8_t __ret;
@@ -65211,48 +55237,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrecpe_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecped_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrecped_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpes_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrecpes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -65270,263 +55269,180 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecpxd_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrecpxd_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpxs_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrecpxs_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
   __ret; \
 })
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u32(__p0_212, __p1_212, __p2_212) __extension__ ({ \
+  uint16x4_t __s0_212 = __p0_212; \
+  uint32x4_t __s1_212 = __p1_212; \
+  uint16x8_t __ret_212; \
+  __ret_212 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_212), (uint16x4_t)(vrshrn_n_u32(__s1_212, __p2_212)))); \
+  __ret_212; \
+})
 #else
-#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
-  __ret; \
+#define vrshrn_high_n_u32(__p0_213, __p1_213, __p2_213) __extension__ ({ \
+  uint16x4_t __s0_213 = __p0_213; \
+  uint32x4_t __s1_213 = __p1_213; \
+  uint16x4_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 3, 2, 1, 0); \
+  uint32x4_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 3, 2, 1, 0); \
+  uint16x8_t __ret_213; \
+  __ret_213 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_213), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_213, __p2_213)))); \
+  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_213; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u32(__p0_218, __p1_218, __p2_218) __extension__ ({ \
-  uint16x4_t __s0_218 = __p0_218; \
-  uint32x4_t __s1_218 = __p1_218; \
-  uint16x8_t __ret_218; \
-  __ret_218 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_218), (uint16x4_t)(vrshrn_n_u32(__s1_218, __p2_218)))); \
+#define vrshrn_high_n_u64(__p0_214, __p1_214, __p2_214) __extension__ ({ \
+  uint32x2_t __s0_214 = __p0_214; \
+  uint64x2_t __s1_214 = __p1_214; \
+  uint32x4_t __ret_214; \
+  __ret_214 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_214), (uint32x2_t)(vrshrn_n_u64(__s1_214, __p2_214)))); \
+  __ret_214; \
+})
+#else
+#define vrshrn_high_n_u64(__p0_215, __p1_215, __p2_215) __extension__ ({ \
+  uint32x2_t __s0_215 = __p0_215; \
+  uint64x2_t __s1_215 = __p1_215; \
+  uint32x2_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 1, 0); \
+  uint64x2_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 1, 0); \
+  uint32x4_t __ret_215; \
+  __ret_215 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_215), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_215, __p2_215)))); \
+  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 3, 2, 1, 0); \
+  __ret_215; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u16(__p0_216, __p1_216, __p2_216) __extension__ ({ \
+  uint8x8_t __s0_216 = __p0_216; \
+  uint16x8_t __s1_216 = __p1_216; \
+  uint8x16_t __ret_216; \
+  __ret_216 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_216), (uint8x8_t)(vrshrn_n_u16(__s1_216, __p2_216)))); \
+  __ret_216; \
+})
+#else
+#define vrshrn_high_n_u16(__p0_217, __p1_217, __p2_217) __extension__ ({ \
+  uint8x8_t __s0_217 = __p0_217; \
+  uint16x8_t __s1_217 = __p1_217; \
+  uint8x8_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_217; \
+  __ret_217 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_217), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_217, __p2_217)))); \
+  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_217; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_s32(__p0_218, __p1_218, __p2_218) __extension__ ({ \
+  int16x4_t __s0_218 = __p0_218; \
+  int32x4_t __s1_218 = __p1_218; \
+  int16x8_t __ret_218; \
+  __ret_218 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_218), (int16x4_t)(vrshrn_n_s32(__s1_218, __p2_218)))); \
   __ret_218; \
 })
 #else
-#define vrshrn_high_n_u32(__p0_219, __p1_219, __p2_219) __extension__ ({ \
-  uint16x4_t __s0_219 = __p0_219; \
-  uint32x4_t __s1_219 = __p1_219; \
-  uint16x4_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 3, 2, 1, 0); \
-  uint32x4_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 3, 2, 1, 0); \
-  uint16x8_t __ret_219; \
-  __ret_219 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_219), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_219, __p2_219)))); \
+#define vrshrn_high_n_s32(__p0_219, __p1_219, __p2_219) __extension__ ({ \
+  int16x4_t __s0_219 = __p0_219; \
+  int32x4_t __s1_219 = __p1_219; \
+  int16x4_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 3, 2, 1, 0); \
+  int32x4_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 3, 2, 1, 0); \
+  int16x8_t __ret_219; \
+  __ret_219 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_219), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_219, __p2_219)))); \
   __ret_219 = __builtin_shufflevector(__ret_219, __ret_219, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_219; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u64(__p0_220, __p1_220, __p2_220) __extension__ ({ \
-  uint32x2_t __s0_220 = __p0_220; \
-  uint64x2_t __s1_220 = __p1_220; \
-  uint32x4_t __ret_220; \
-  __ret_220 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_220), (uint32x2_t)(vrshrn_n_u64(__s1_220, __p2_220)))); \
+#define vrshrn_high_n_s64(__p0_220, __p1_220, __p2_220) __extension__ ({ \
+  int32x2_t __s0_220 = __p0_220; \
+  int64x2_t __s1_220 = __p1_220; \
+  int32x4_t __ret_220; \
+  __ret_220 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_220), (int32x2_t)(vrshrn_n_s64(__s1_220, __p2_220)))); \
   __ret_220; \
 })
 #else
-#define vrshrn_high_n_u64(__p0_221, __p1_221, __p2_221) __extension__ ({ \
-  uint32x2_t __s0_221 = __p0_221; \
-  uint64x2_t __s1_221 = __p1_221; \
-  uint32x2_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 1, 0); \
-  uint64x2_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 1, 0); \
-  uint32x4_t __ret_221; \
-  __ret_221 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_221), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_221, __p2_221)))); \
+#define vrshrn_high_n_s64(__p0_221, __p1_221, __p2_221) __extension__ ({ \
+  int32x2_t __s0_221 = __p0_221; \
+  int64x2_t __s1_221 = __p1_221; \
+  int32x2_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 1, 0); \
+  int64x2_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 1, 0); \
+  int32x4_t __ret_221; \
+  __ret_221 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_221), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_221, __p2_221)))); \
   __ret_221 = __builtin_shufflevector(__ret_221, __ret_221, 3, 2, 1, 0); \
   __ret_221; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
-  uint8x8_t __s0_222 = __p0_222; \
-  uint16x8_t __s1_222 = __p1_222; \
-  uint8x16_t __ret_222; \
-  __ret_222 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_222), (uint8x8_t)(vrshrn_n_u16(__s1_222, __p2_222)))); \
+#define vrshrn_high_n_s16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
+  int8x8_t __s0_222 = __p0_222; \
+  int16x8_t __s1_222 = __p1_222; \
+  int8x16_t __ret_222; \
+  __ret_222 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_222), (int8x8_t)(vrshrn_n_s16(__s1_222, __p2_222)))); \
   __ret_222; \
 })
 #else
-#define vrshrn_high_n_u16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
-  uint8x8_t __s0_223 = __p0_223; \
-  uint16x8_t __s1_223 = __p1_223; \
-  uint8x8_t __rev0_223;  __rev0_223 = __builtin_shufflevector(__s0_223, __s0_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_223; \
-  __ret_223 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_223), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_223, __p2_223)))); \
+#define vrshrn_high_n_s16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
+  int8x8_t __s0_223 = __p0_223; \
+  int16x8_t __s1_223 = __p1_223; \
+  int8x8_t __rev0_223;  __rev0_223 = __builtin_shufflevector(__s0_223, __s0_223, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_223; \
+  __ret_223 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_223), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_223, __p2_223)))); \
   __ret_223 = __builtin_shufflevector(__ret_223, __ret_223, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_223; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s32(__p0_224, __p1_224, __p2_224) __extension__ ({ \
-  int16x4_t __s0_224 = __p0_224; \
-  int32x4_t __s1_224 = __p1_224; \
-  int16x8_t __ret_224; \
-  __ret_224 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_224), (int16x4_t)(vrshrn_n_s32(__s1_224, __p2_224)))); \
-  __ret_224; \
-})
-#else
-#define vrshrn_high_n_s32(__p0_225, __p1_225, __p2_225) __extension__ ({ \
-  int16x4_t __s0_225 = __p0_225; \
-  int32x4_t __s1_225 = __p1_225; \
-  int16x4_t __rev0_225;  __rev0_225 = __builtin_shufflevector(__s0_225, __s0_225, 3, 2, 1, 0); \
-  int32x4_t __rev1_225;  __rev1_225 = __builtin_shufflevector(__s1_225, __s1_225, 3, 2, 1, 0); \
-  int16x8_t __ret_225; \
-  __ret_225 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_225), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_225, __p2_225)))); \
-  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_225; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s64(__p0_226, __p1_226, __p2_226) __extension__ ({ \
-  int32x2_t __s0_226 = __p0_226; \
-  int64x2_t __s1_226 = __p1_226; \
-  int32x4_t __ret_226; \
-  __ret_226 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_226), (int32x2_t)(vrshrn_n_s64(__s1_226, __p2_226)))); \
-  __ret_226; \
-})
-#else
-#define vrshrn_high_n_s64(__p0_227, __p1_227, __p2_227) __extension__ ({ \
-  int32x2_t __s0_227 = __p0_227; \
-  int64x2_t __s1_227 = __p1_227; \
-  int32x2_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 1, 0); \
-  int64x2_t __rev1_227;  __rev1_227 = __builtin_shufflevector(__s1_227, __s1_227, 1, 0); \
-  int32x4_t __ret_227; \
-  __ret_227 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_227), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_227, __p2_227)))); \
-  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 3, 2, 1, 0); \
-  __ret_227; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s16(__p0_228, __p1_228, __p2_228) __extension__ ({ \
-  int8x8_t __s0_228 = __p0_228; \
-  int16x8_t __s1_228 = __p1_228; \
-  int8x16_t __ret_228; \
-  __ret_228 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_228), (int8x8_t)(vrshrn_n_s16(__s1_228, __p2_228)))); \
-  __ret_228; \
-})
-#else
-#define vrshrn_high_n_s16(__p0_229, __p1_229, __p2_229) __extension__ ({ \
-  int8x8_t __s0_229 = __p0_229; \
-  int16x8_t __s1_229 = __p1_229; \
-  int8x8_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_229;  __rev1_229 = __builtin_shufflevector(__s1_229, __s1_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_229; \
-  __ret_229 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_229), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_229, __p2_229)))); \
-  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_229; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 42);
@@ -65542,48 +55458,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrsqrted_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrsqrted_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrsqrtes_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrsqrtes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -65601,49 +55490,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -65651,17 +55512,6 @@
   __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -65669,16 +55519,6 @@
   __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -65787,7 +55627,6 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -65795,23 +55634,6 @@
   __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64_t __s0 = __p0; \
@@ -65866,7 +55688,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -65874,350 +55695,274 @@
   __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vshld_n_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vshld_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
   __ret; \
 })
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u8(__p0_224, __p1_224) __extension__ ({ \
+  uint8x16_t __s0_224 = __p0_224; \
+  uint16x8_t __ret_224; \
+  __ret_224 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_224), __p1_224)); \
+  __ret_224; \
+})
 #else
-#define vshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
-  __ret; \
+#define vshll_high_n_u8(__p0_225, __p1_225) __extension__ ({ \
+  uint8x16_t __s0_225 = __p0_225; \
+  uint8x16_t __rev0_225;  __rev0_225 = __builtin_shufflevector(__s0_225, __s0_225, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_225; \
+  __ret_225 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_225), __p1_225)); \
+  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_225; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u8(__p0_230, __p1_230) __extension__ ({ \
-  uint8x16_t __s0_230 = __p0_230; \
-  uint16x8_t __ret_230; \
-  __ret_230 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_230), __p1_230)); \
+#define vshll_high_n_u32(__p0_226, __p1_226) __extension__ ({ \
+  uint32x4_t __s0_226 = __p0_226; \
+  uint64x2_t __ret_226; \
+  __ret_226 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_226), __p1_226)); \
+  __ret_226; \
+})
+#else
+#define vshll_high_n_u32(__p0_227, __p1_227) __extension__ ({ \
+  uint32x4_t __s0_227 = __p0_227; \
+  uint32x4_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 3, 2, 1, 0); \
+  uint64x2_t __ret_227; \
+  __ret_227 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_227), __p1_227)); \
+  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 1, 0); \
+  __ret_227; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u16(__p0_228, __p1_228) __extension__ ({ \
+  uint16x8_t __s0_228 = __p0_228; \
+  uint32x4_t __ret_228; \
+  __ret_228 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_228), __p1_228)); \
+  __ret_228; \
+})
+#else
+#define vshll_high_n_u16(__p0_229, __p1_229) __extension__ ({ \
+  uint16x8_t __s0_229 = __p0_229; \
+  uint16x8_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret_229; \
+  __ret_229 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_229), __p1_229)); \
+  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 3, 2, 1, 0); \
+  __ret_229; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_s8(__p0_230, __p1_230) __extension__ ({ \
+  int8x16_t __s0_230 = __p0_230; \
+  int16x8_t __ret_230; \
+  __ret_230 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_230), __p1_230)); \
   __ret_230; \
 })
 #else
-#define vshll_high_n_u8(__p0_231, __p1_231) __extension__ ({ \
-  uint8x16_t __s0_231 = __p0_231; \
-  uint8x16_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_231; \
-  __ret_231 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_231), __p1_231)); \
+#define vshll_high_n_s8(__p0_231, __p1_231) __extension__ ({ \
+  int8x16_t __s0_231 = __p0_231; \
+  int8x16_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_231; \
+  __ret_231 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_231), __p1_231)); \
   __ret_231 = __builtin_shufflevector(__ret_231, __ret_231, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_231; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u32(__p0_232, __p1_232) __extension__ ({ \
-  uint32x4_t __s0_232 = __p0_232; \
-  uint64x2_t __ret_232; \
-  __ret_232 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_232), __p1_232)); \
+#define vshll_high_n_s32(__p0_232, __p1_232) __extension__ ({ \
+  int32x4_t __s0_232 = __p0_232; \
+  int64x2_t __ret_232; \
+  __ret_232 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_232), __p1_232)); \
   __ret_232; \
 })
 #else
-#define vshll_high_n_u32(__p0_233, __p1_233) __extension__ ({ \
-  uint32x4_t __s0_233 = __p0_233; \
-  uint32x4_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 3, 2, 1, 0); \
-  uint64x2_t __ret_233; \
-  __ret_233 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_233), __p1_233)); \
+#define vshll_high_n_s32(__p0_233, __p1_233) __extension__ ({ \
+  int32x4_t __s0_233 = __p0_233; \
+  int32x4_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 3, 2, 1, 0); \
+  int64x2_t __ret_233; \
+  __ret_233 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_233), __p1_233)); \
   __ret_233 = __builtin_shufflevector(__ret_233, __ret_233, 1, 0); \
   __ret_233; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u16(__p0_234, __p1_234) __extension__ ({ \
-  uint16x8_t __s0_234 = __p0_234; \
-  uint32x4_t __ret_234; \
-  __ret_234 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_234), __p1_234)); \
+#define vshll_high_n_s16(__p0_234, __p1_234) __extension__ ({ \
+  int16x8_t __s0_234 = __p0_234; \
+  int32x4_t __ret_234; \
+  __ret_234 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_234), __p1_234)); \
   __ret_234; \
 })
 #else
-#define vshll_high_n_u16(__p0_235, __p1_235) __extension__ ({ \
-  uint16x8_t __s0_235 = __p0_235; \
-  uint16x8_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret_235; \
-  __ret_235 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_235), __p1_235)); \
+#define vshll_high_n_s16(__p0_235, __p1_235) __extension__ ({ \
+  int16x8_t __s0_235 = __p0_235; \
+  int16x8_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret_235; \
+  __ret_235 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_235), __p1_235)); \
   __ret_235 = __builtin_shufflevector(__ret_235, __ret_235, 3, 2, 1, 0); \
   __ret_235; \
 })
 #endif
 
+#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s8(__p0_236, __p1_236) __extension__ ({ \
-  int8x16_t __s0_236 = __p0_236; \
-  int16x8_t __ret_236; \
-  __ret_236 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_236), __p1_236)); \
+#define vshrn_high_n_u32(__p0_236, __p1_236, __p2_236) __extension__ ({ \
+  uint16x4_t __s0_236 = __p0_236; \
+  uint32x4_t __s1_236 = __p1_236; \
+  uint16x8_t __ret_236; \
+  __ret_236 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_236), (uint16x4_t)(vshrn_n_u32(__s1_236, __p2_236)))); \
   __ret_236; \
 })
 #else
-#define vshll_high_n_s8(__p0_237, __p1_237) __extension__ ({ \
-  int8x16_t __s0_237 = __p0_237; \
-  int8x16_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_237; \
-  __ret_237 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_237), __p1_237)); \
+#define vshrn_high_n_u32(__p0_237, __p1_237, __p2_237) __extension__ ({ \
+  uint16x4_t __s0_237 = __p0_237; \
+  uint32x4_t __s1_237 = __p1_237; \
+  uint16x4_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 3, 2, 1, 0); \
+  uint32x4_t __rev1_237;  __rev1_237 = __builtin_shufflevector(__s1_237, __s1_237, 3, 2, 1, 0); \
+  uint16x8_t __ret_237; \
+  __ret_237 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_237), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_237, __p2_237)))); \
   __ret_237 = __builtin_shufflevector(__ret_237, __ret_237, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_237; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s32(__p0_238, __p1_238) __extension__ ({ \
-  int32x4_t __s0_238 = __p0_238; \
-  int64x2_t __ret_238; \
-  __ret_238 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_238), __p1_238)); \
+#define vshrn_high_n_u64(__p0_238, __p1_238, __p2_238) __extension__ ({ \
+  uint32x2_t __s0_238 = __p0_238; \
+  uint64x2_t __s1_238 = __p1_238; \
+  uint32x4_t __ret_238; \
+  __ret_238 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_238), (uint32x2_t)(vshrn_n_u64(__s1_238, __p2_238)))); \
   __ret_238; \
 })
 #else
-#define vshll_high_n_s32(__p0_239, __p1_239) __extension__ ({ \
-  int32x4_t __s0_239 = __p0_239; \
-  int32x4_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 3, 2, 1, 0); \
-  int64x2_t __ret_239; \
-  __ret_239 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_239), __p1_239)); \
-  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 1, 0); \
+#define vshrn_high_n_u64(__p0_239, __p1_239, __p2_239) __extension__ ({ \
+  uint32x2_t __s0_239 = __p0_239; \
+  uint64x2_t __s1_239 = __p1_239; \
+  uint32x2_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 1, 0); \
+  uint64x2_t __rev1_239;  __rev1_239 = __builtin_shufflevector(__s1_239, __s1_239, 1, 0); \
+  uint32x4_t __ret_239; \
+  __ret_239 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_239), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_239, __p2_239)))); \
+  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 3, 2, 1, 0); \
   __ret_239; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s16(__p0_240, __p1_240) __extension__ ({ \
-  int16x8_t __s0_240 = __p0_240; \
-  int32x4_t __ret_240; \
-  __ret_240 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_240), __p1_240)); \
+#define vshrn_high_n_u16(__p0_240, __p1_240, __p2_240) __extension__ ({ \
+  uint8x8_t __s0_240 = __p0_240; \
+  uint16x8_t __s1_240 = __p1_240; \
+  uint8x16_t __ret_240; \
+  __ret_240 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_240), (uint8x8_t)(vshrn_n_u16(__s1_240, __p2_240)))); \
   __ret_240; \
 })
 #else
-#define vshll_high_n_s16(__p0_241, __p1_241) __extension__ ({ \
-  int16x8_t __s0_241 = __p0_241; \
-  int16x8_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret_241; \
-  __ret_241 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_241), __p1_241)); \
-  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 3, 2, 1, 0); \
+#define vshrn_high_n_u16(__p0_241, __p1_241, __p2_241) __extension__ ({ \
+  uint8x8_t __s0_241 = __p0_241; \
+  uint16x8_t __s1_241 = __p1_241; \
+  uint8x8_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_241;  __rev1_241 = __builtin_shufflevector(__s1_241, __s1_241, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_241; \
+  __ret_241 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_241), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_241, __p2_241)))); \
+  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_241; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
-  uint16x4_t __s0_242 = __p0_242; \
-  uint32x4_t __s1_242 = __p1_242; \
-  uint16x8_t __ret_242; \
-  __ret_242 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_242), (uint16x4_t)(vshrn_n_u32(__s1_242, __p2_242)))); \
+#define vshrn_high_n_s32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
+  int16x4_t __s0_242 = __p0_242; \
+  int32x4_t __s1_242 = __p1_242; \
+  int16x8_t __ret_242; \
+  __ret_242 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_242), (int16x4_t)(vshrn_n_s32(__s1_242, __p2_242)))); \
   __ret_242; \
 })
 #else
-#define vshrn_high_n_u32(__p0_243, __p1_243, __p2_243) __extension__ ({ \
-  uint16x4_t __s0_243 = __p0_243; \
-  uint32x4_t __s1_243 = __p1_243; \
-  uint16x4_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 3, 2, 1, 0); \
-  uint32x4_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 3, 2, 1, 0); \
-  uint16x8_t __ret_243; \
-  __ret_243 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_243), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_243, __p2_243)))); \
+#define vshrn_high_n_s32(__p0_243, __p1_243, __p2_243) __extension__ ({ \
+  int16x4_t __s0_243 = __p0_243; \
+  int32x4_t __s1_243 = __p1_243; \
+  int16x4_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 3, 2, 1, 0); \
+  int32x4_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 3, 2, 1, 0); \
+  int16x8_t __ret_243; \
+  __ret_243 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_243), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_243, __p2_243)))); \
   __ret_243 = __builtin_shufflevector(__ret_243, __ret_243, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_243; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u64(__p0_244, __p1_244, __p2_244) __extension__ ({ \
-  uint32x2_t __s0_244 = __p0_244; \
-  uint64x2_t __s1_244 = __p1_244; \
-  uint32x4_t __ret_244; \
-  __ret_244 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_244), (uint32x2_t)(vshrn_n_u64(__s1_244, __p2_244)))); \
+#define vshrn_high_n_s64(__p0_244, __p1_244, __p2_244) __extension__ ({ \
+  int32x2_t __s0_244 = __p0_244; \
+  int64x2_t __s1_244 = __p1_244; \
+  int32x4_t __ret_244; \
+  __ret_244 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_244), (int32x2_t)(vshrn_n_s64(__s1_244, __p2_244)))); \
   __ret_244; \
 })
 #else
-#define vshrn_high_n_u64(__p0_245, __p1_245, __p2_245) __extension__ ({ \
-  uint32x2_t __s0_245 = __p0_245; \
-  uint64x2_t __s1_245 = __p1_245; \
-  uint32x2_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 1, 0); \
-  uint64x2_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 1, 0); \
-  uint32x4_t __ret_245; \
-  __ret_245 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_245), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_245, __p2_245)))); \
+#define vshrn_high_n_s64(__p0_245, __p1_245, __p2_245) __extension__ ({ \
+  int32x2_t __s0_245 = __p0_245; \
+  int64x2_t __s1_245 = __p1_245; \
+  int32x2_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 1, 0); \
+  int64x2_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 1, 0); \
+  int32x4_t __ret_245; \
+  __ret_245 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_245), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_245, __p2_245)))); \
   __ret_245 = __builtin_shufflevector(__ret_245, __ret_245, 3, 2, 1, 0); \
   __ret_245; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u16(__p0_246, __p1_246, __p2_246) __extension__ ({ \
-  uint8x8_t __s0_246 = __p0_246; \
-  uint16x8_t __s1_246 = __p1_246; \
-  uint8x16_t __ret_246; \
-  __ret_246 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_246), (uint8x8_t)(vshrn_n_u16(__s1_246, __p2_246)))); \
+#define vshrn_high_n_s16(__p0_246, __p1_246, __p2_246) __extension__ ({ \
+  int8x8_t __s0_246 = __p0_246; \
+  int16x8_t __s1_246 = __p1_246; \
+  int8x16_t __ret_246; \
+  __ret_246 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_246), (int8x8_t)(vshrn_n_s16(__s1_246, __p2_246)))); \
   __ret_246; \
 })
 #else
-#define vshrn_high_n_u16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
-  uint8x8_t __s0_247 = __p0_247; \
-  uint16x8_t __s1_247 = __p1_247; \
-  uint8x8_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_247; \
-  __ret_247 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_247), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_247, __p2_247)))); \
+#define vshrn_high_n_s16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
+  int8x8_t __s0_247 = __p0_247; \
+  int16x8_t __s1_247 = __p1_247; \
+  int8x8_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_247; \
+  __ret_247 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_247), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_247, __p2_247)))); \
   __ret_247 = __builtin_shufflevector(__ret_247, __ret_247, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_247; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s32(__p0_248, __p1_248, __p2_248) __extension__ ({ \
-  int16x4_t __s0_248 = __p0_248; \
-  int32x4_t __s1_248 = __p1_248; \
-  int16x8_t __ret_248; \
-  __ret_248 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_248), (int16x4_t)(vshrn_n_s32(__s1_248, __p2_248)))); \
-  __ret_248; \
-})
-#else
-#define vshrn_high_n_s32(__p0_249, __p1_249, __p2_249) __extension__ ({ \
-  int16x4_t __s0_249 = __p0_249; \
-  int32x4_t __s1_249 = __p1_249; \
-  int16x4_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 3, 2, 1, 0); \
-  int32x4_t __rev1_249;  __rev1_249 = __builtin_shufflevector(__s1_249, __s1_249, 3, 2, 1, 0); \
-  int16x8_t __ret_249; \
-  __ret_249 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_249), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_249, __p2_249)))); \
-  __ret_249 = __builtin_shufflevector(__ret_249, __ret_249, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_249; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s64(__p0_250, __p1_250, __p2_250) __extension__ ({ \
-  int32x2_t __s0_250 = __p0_250; \
-  int64x2_t __s1_250 = __p1_250; \
-  int32x4_t __ret_250; \
-  __ret_250 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_250), (int32x2_t)(vshrn_n_s64(__s1_250, __p2_250)))); \
-  __ret_250; \
-})
-#else
-#define vshrn_high_n_s64(__p0_251, __p1_251, __p2_251) __extension__ ({ \
-  int32x2_t __s0_251 = __p0_251; \
-  int64x2_t __s1_251 = __p1_251; \
-  int32x2_t __rev0_251;  __rev0_251 = __builtin_shufflevector(__s0_251, __s0_251, 1, 0); \
-  int64x2_t __rev1_251;  __rev1_251 = __builtin_shufflevector(__s1_251, __s1_251, 1, 0); \
-  int32x4_t __ret_251; \
-  __ret_251 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_251), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_251, __p2_251)))); \
-  __ret_251 = __builtin_shufflevector(__ret_251, __ret_251, 3, 2, 1, 0); \
-  __ret_251; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s16(__p0_252, __p1_252, __p2_252) __extension__ ({ \
-  int8x8_t __s0_252 = __p0_252; \
-  int16x8_t __s1_252 = __p1_252; \
-  int8x16_t __ret_252; \
-  __ret_252 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_252), (int8x8_t)(vshrn_n_s16(__s1_252, __p2_252)))); \
-  __ret_252; \
-})
-#else
-#define vshrn_high_n_s16(__p0_253, __p1_253, __p2_253) __extension__ ({ \
-  int8x8_t __s0_253 = __p0_253; \
-  int16x8_t __s1_253 = __p1_253; \
-  int8x8_t __rev0_253;  __rev0_253 = __builtin_shufflevector(__s0_253, __s0_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_253;  __rev1_253 = __builtin_shufflevector(__s1_253, __s1_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_253; \
-  __ret_253 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_253), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_253, __p2_253)))); \
-  __ret_253 = __builtin_shufflevector(__ret_253, __ret_253, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_253; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66225,17 +55970,6 @@
   __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66243,17 +55977,6 @@
   __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -66261,16 +55984,6 @@
   __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -66292,62 +56005,26 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
   uint8x16_t __ret;
@@ -66450,20 +56127,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -66513,20 +56181,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vsqrt_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vsqrt_f32(float32x2_t __p0) {
   float32x2_t __ret;
@@ -66543,7 +56202,6 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66551,17 +56209,6 @@
   __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66569,17 +56216,6 @@
   __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66587,17 +56223,6 @@
   __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66605,17 +56230,6 @@
   __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -66623,16 +56237,6 @@
   __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -66654,18 +56258,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
 })
-#else
-#define vst1_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64(__p0, __p1) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -66692,30 +56288,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
 })
-#else
-#define vst1_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
 })
-#else
-#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -66742,30 +56322,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
 })
-#else
-#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x2(__p0, __p1) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
 })
-#else
-#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
   poly64x2x2_t __s1 = __p1; \
@@ -66796,30 +56360,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x2(__p0, __p1) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
 })
-#else
-#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x3(__p0, __p1) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
 })
-#else
-#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
   poly64x2x3_t __s1 = __p1; \
@@ -66852,30 +56400,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x3(__p0, __p1) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
 })
-#else
-#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x4(__p0, __p1) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
 })
-#else
-#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
   poly64x2x4_t __s1 = __p1; \
@@ -66910,30 +56442,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x4(__p0, __p1) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
 })
-#else
-#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_p64(__p0, __p1) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
 })
-#else
-#define vst2_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x2_t __s1 = __p1; \
@@ -66994,30 +56510,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_f64(__p0, __p1) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
 })
-#else
-#define vst2_f64(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
 })
-#else
-#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x2_t __s1 = __p1; \
@@ -67123,54 +56623,22 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
 })
-#else
-#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
 })
-#else
-#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
 })
-#else
-#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_p64(__p0, __p1) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
 })
-#else
-#define vst3_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x3_t __s1 = __p1; \
@@ -67235,30 +56703,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_f64(__p0, __p1) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
 })
-#else
-#define vst3_f64(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
 })
-#else
-#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x3_t __s1 = __p1; \
@@ -67371,54 +56823,22 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
 })
-#else
-#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
 })
-#else
-#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
 })
-#else
-#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_p64(__p0, __p1) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
 })
-#else
-#define vst4_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x4_t __s1 = __p1; \
@@ -67487,30 +56907,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_f64(__p0, __p1) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
 })
-#else
-#define vst4_f64(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
 })
-#else
-#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x4_t __s1 = __p1; \
@@ -67630,82 +57034,32 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
 })
-#else
-#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
 })
-#else
-#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
 })
-#else
-#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vstrq_p128(__p0, __p1) __extension__ ({ \
   poly128_t __s1 = __p1; \
   __builtin_neon_vstrq_p128(__p0, __s1); \
 })
-#else
-#define vstrq_p128(__p0, __p1) __extension__ ({ \
-  poly128_t __s1 = __p1; \
-  __builtin_neon_vstrq_p128(__p0, __s1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -67723,20 +57077,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -68797,20 +58142,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
   uint64x2_t __ret;
@@ -68862,118 +58198,46 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
   int8x16_t __ret;
@@ -69076,20 +58340,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
   int16x4_t __ret;
@@ -71157,60 +60412,60 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vget_lane_f16(__p0_254, __p1_254) __extension__ ({ \
-  float16x4_t __s0_254 = __p0_254; \
-  float16_t __ret_254; \
-float16x4_t __reint_254 = __s0_254; \
-int16_t __reint1_254 = vget_lane_s16(*(int16x4_t *) &__reint_254, __p1_254); \
-  __ret_254 = *(float16_t *) &__reint1_254; \
-  __ret_254; \
+#define vget_lane_f16(__p0_248, __p1_248) __extension__ ({ \
+  float16x4_t __s0_248 = __p0_248; \
+  float16_t __ret_248; \
+float16x4_t __reint_248 = __s0_248; \
+int16_t __reint1_248 = vget_lane_s16(*(int16x4_t *) &__reint_248, __p1_248); \
+  __ret_248 = *(float16_t *) &__reint1_248; \
+  __ret_248; \
 })
 #else
-#define vget_lane_f16(__p0_255, __p1_255) __extension__ ({ \
-  float16x4_t __s0_255 = __p0_255; \
-  float16x4_t __rev0_255;  __rev0_255 = __builtin_shufflevector(__s0_255, __s0_255, 3, 2, 1, 0); \
-  float16_t __ret_255; \
-float16x4_t __reint_255 = __rev0_255; \
-int16_t __reint1_255 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_255, __p1_255); \
-  __ret_255 = *(float16_t *) &__reint1_255; \
-  __ret_255; \
+#define vget_lane_f16(__p0_249, __p1_249) __extension__ ({ \
+  float16x4_t __s0_249 = __p0_249; \
+  float16x4_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 3, 2, 1, 0); \
+  float16_t __ret_249; \
+float16x4_t __reint_249 = __rev0_249; \
+int16_t __reint1_249 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_249, __p1_249); \
+  __ret_249 = *(float16_t *) &__reint1_249; \
+  __ret_249; \
 })
-#define __noswap_vget_lane_f16(__p0_256, __p1_256) __extension__ ({ \
-  float16x4_t __s0_256 = __p0_256; \
-  float16_t __ret_256; \
-float16x4_t __reint_256 = __s0_256; \
-int16_t __reint1_256 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_256, __p1_256); \
-  __ret_256 = *(float16_t *) &__reint1_256; \
-  __ret_256; \
+#define __noswap_vget_lane_f16(__p0_250, __p1_250) __extension__ ({ \
+  float16x4_t __s0_250 = __p0_250; \
+  float16_t __ret_250; \
+float16x4_t __reint_250 = __s0_250; \
+int16_t __reint1_250 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_250, __p1_250); \
+  __ret_250 = *(float16_t *) &__reint1_250; \
+  __ret_250; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_f16(__p0_257, __p1_257) __extension__ ({ \
-  float16x8_t __s0_257 = __p0_257; \
-  float16_t __ret_257; \
-float16x8_t __reint_257 = __s0_257; \
-int16_t __reint1_257 = vgetq_lane_s16(*(int16x8_t *) &__reint_257, __p1_257); \
-  __ret_257 = *(float16_t *) &__reint1_257; \
-  __ret_257; \
+#define vgetq_lane_f16(__p0_251, __p1_251) __extension__ ({ \
+  float16x8_t __s0_251 = __p0_251; \
+  float16_t __ret_251; \
+float16x8_t __reint_251 = __s0_251; \
+int16_t __reint1_251 = vgetq_lane_s16(*(int16x8_t *) &__reint_251, __p1_251); \
+  __ret_251 = *(float16_t *) &__reint1_251; \
+  __ret_251; \
 })
 #else
-#define vgetq_lane_f16(__p0_258, __p1_258) __extension__ ({ \
-  float16x8_t __s0_258 = __p0_258; \
-  float16x8_t __rev0_258;  __rev0_258 = __builtin_shufflevector(__s0_258, __s0_258, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_258; \
-float16x8_t __reint_258 = __rev0_258; \
-int16_t __reint1_258 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_258, __p1_258); \
-  __ret_258 = *(float16_t *) &__reint1_258; \
-  __ret_258; \
+#define vgetq_lane_f16(__p0_252, __p1_252) __extension__ ({ \
+  float16x8_t __s0_252 = __p0_252; \
+  float16x8_t __rev0_252;  __rev0_252 = __builtin_shufflevector(__s0_252, __s0_252, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16_t __ret_252; \
+float16x8_t __reint_252 = __rev0_252; \
+int16_t __reint1_252 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_252, __p1_252); \
+  __ret_252 = *(float16_t *) &__reint1_252; \
+  __ret_252; \
 })
-#define __noswap_vgetq_lane_f16(__p0_259, __p1_259) __extension__ ({ \
-  float16x8_t __s0_259 = __p0_259; \
-  float16_t __ret_259; \
-float16x8_t __reint_259 = __s0_259; \
-int16_t __reint1_259 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_259, __p1_259); \
-  __ret_259 = *(float16_t *) &__reint1_259; \
-  __ret_259; \
+#define __noswap_vgetq_lane_f16(__p0_253, __p1_253) __extension__ ({ \
+  float16x8_t __s0_253 = __p0_253; \
+  float16_t __ret_253; \
+float16x8_t __reint_253 = __s0_253; \
+int16_t __reint1_253 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_253, __p1_253); \
+  __ret_253 = *(float16_t *) &__reint1_253; \
+  __ret_253; \
 })
 #endif
 
@@ -71859,708 +61114,672 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vset_lane_f16(__p0_260, __p1_260, __p2_260) __extension__ ({ \
-  float16_t __s0_260 = __p0_260; \
-  float16x4_t __s1_260 = __p1_260; \
-  float16x4_t __ret_260; \
-float16_t __reint_260 = __s0_260; \
-float16x4_t __reint1_260 = __s1_260; \
-int16x4_t __reint2_260 = vset_lane_s16(*(int16_t *) &__reint_260, *(int16x4_t *) &__reint1_260, __p2_260); \
-  __ret_260 = *(float16x4_t *) &__reint2_260; \
-  __ret_260; \
+#define vset_lane_f16(__p0_254, __p1_254, __p2_254) __extension__ ({ \
+  float16_t __s0_254 = __p0_254; \
+  float16x4_t __s1_254 = __p1_254; \
+  float16x4_t __ret_254; \
+float16_t __reint_254 = __s0_254; \
+float16x4_t __reint1_254 = __s1_254; \
+int16x4_t __reint2_254 = vset_lane_s16(*(int16_t *) &__reint_254, *(int16x4_t *) &__reint1_254, __p2_254); \
+  __ret_254 = *(float16x4_t *) &__reint2_254; \
+  __ret_254; \
 })
 #else
-#define vset_lane_f16(__p0_261, __p1_261, __p2_261) __extension__ ({ \
-  float16_t __s0_261 = __p0_261; \
-  float16x4_t __s1_261 = __p1_261; \
-  float16x4_t __rev1_261;  __rev1_261 = __builtin_shufflevector(__s1_261, __s1_261, 3, 2, 1, 0); \
-  float16x4_t __ret_261; \
-float16_t __reint_261 = __s0_261; \
-float16x4_t __reint1_261 = __rev1_261; \
-int16x4_t __reint2_261 = __noswap_vset_lane_s16(*(int16_t *) &__reint_261, *(int16x4_t *) &__reint1_261, __p2_261); \
-  __ret_261 = *(float16x4_t *) &__reint2_261; \
-  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 3, 2, 1, 0); \
-  __ret_261; \
+#define vset_lane_f16(__p0_255, __p1_255, __p2_255) __extension__ ({ \
+  float16_t __s0_255 = __p0_255; \
+  float16x4_t __s1_255 = __p1_255; \
+  float16x4_t __rev1_255;  __rev1_255 = __builtin_shufflevector(__s1_255, __s1_255, 3, 2, 1, 0); \
+  float16x4_t __ret_255; \
+float16_t __reint_255 = __s0_255; \
+float16x4_t __reint1_255 = __rev1_255; \
+int16x4_t __reint2_255 = __noswap_vset_lane_s16(*(int16_t *) &__reint_255, *(int16x4_t *) &__reint1_255, __p2_255); \
+  __ret_255 = *(float16x4_t *) &__reint2_255; \
+  __ret_255 = __builtin_shufflevector(__ret_255, __ret_255, 3, 2, 1, 0); \
+  __ret_255; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_f16(__p0_262, __p1_262, __p2_262) __extension__ ({ \
-  float16_t __s0_262 = __p0_262; \
-  float16x8_t __s1_262 = __p1_262; \
-  float16x8_t __ret_262; \
-float16_t __reint_262 = __s0_262; \
-float16x8_t __reint1_262 = __s1_262; \
-int16x8_t __reint2_262 = vsetq_lane_s16(*(int16_t *) &__reint_262, *(int16x8_t *) &__reint1_262, __p2_262); \
-  __ret_262 = *(float16x8_t *) &__reint2_262; \
-  __ret_262; \
+#define vsetq_lane_f16(__p0_256, __p1_256, __p2_256) __extension__ ({ \
+  float16_t __s0_256 = __p0_256; \
+  float16x8_t __s1_256 = __p1_256; \
+  float16x8_t __ret_256; \
+float16_t __reint_256 = __s0_256; \
+float16x8_t __reint1_256 = __s1_256; \
+int16x8_t __reint2_256 = vsetq_lane_s16(*(int16_t *) &__reint_256, *(int16x8_t *) &__reint1_256, __p2_256); \
+  __ret_256 = *(float16x8_t *) &__reint2_256; \
+  __ret_256; \
 })
 #else
-#define vsetq_lane_f16(__p0_263, __p1_263, __p2_263) __extension__ ({ \
-  float16_t __s0_263 = __p0_263; \
-  float16x8_t __s1_263 = __p1_263; \
-  float16x8_t __rev1_263;  __rev1_263 = __builtin_shufflevector(__s1_263, __s1_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret_263; \
-float16_t __reint_263 = __s0_263; \
-float16x8_t __reint1_263 = __rev1_263; \
-int16x8_t __reint2_263 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_263, *(int16x8_t *) &__reint1_263, __p2_263); \
-  __ret_263 = *(float16x8_t *) &__reint2_263; \
-  __ret_263 = __builtin_shufflevector(__ret_263, __ret_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_263; \
+#define vsetq_lane_f16(__p0_257, __p1_257, __p2_257) __extension__ ({ \
+  float16_t __s0_257 = __p0_257; \
+  float16x8_t __s1_257 = __p1_257; \
+  float16x8_t __rev1_257;  __rev1_257 = __builtin_shufflevector(__s1_257, __s1_257, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __ret_257; \
+float16_t __reint_257 = __s0_257; \
+float16x8_t __reint1_257 = __rev1_257; \
+int16x8_t __reint2_257 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_257, *(int16x8_t *) &__reint1_257, __p2_257); \
+  __ret_257 = *(float16x8_t *) &__reint2_257; \
+  __ret_257 = __builtin_shufflevector(__ret_257, __ret_257, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_257; \
 })
 #endif
 
 #if defined(__ARM_FEATURE_FP16FML) && defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_high_f16(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
-  float32x4_t __s0_264 = __p0_264; \
-  float16x8_t __s1_264 = __p1_264; \
+#define vfmlalq_lane_high_f16(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \
+  float32x4_t __s0_258 = __p0_258; \
+  float16x8_t __s1_258 = __p1_258; \
+  float16x4_t __s2_258 = __p2_258; \
+  float32x4_t __ret_258; \
+  __ret_258 = vfmlalq_high_f16(__s0_258, __s1_258, (float16x8_t) {vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258)}); \
+  __ret_258; \
+})
+#else
+#define vfmlalq_lane_high_f16(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \
+  float32x4_t __s0_259 = __p0_259; \
+  float16x8_t __s1_259 = __p1_259; \
+  float16x4_t __s2_259 = __p2_259; \
+  float32x4_t __rev0_259;  __rev0_259 = __builtin_shufflevector(__s0_259, __s0_259, 3, 2, 1, 0); \
+  float16x8_t __rev1_259;  __rev1_259 = __builtin_shufflevector(__s1_259, __s1_259, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_259;  __rev2_259 = __builtin_shufflevector(__s2_259, __s2_259, 3, 2, 1, 0); \
+  float32x4_t __ret_259; \
+  __ret_259 = __noswap_vfmlalq_high_f16(__rev0_259, __rev1_259, (float16x8_t) {__noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259)}); \
+  __ret_259 = __builtin_shufflevector(__ret_259, __ret_259, 3, 2, 1, 0); \
+  __ret_259; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlal_lane_high_f16(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \
+  float32x2_t __s0_260 = __p0_260; \
+  float16x4_t __s1_260 = __p1_260; \
+  float16x4_t __s2_260 = __p2_260; \
+  float32x2_t __ret_260; \
+  __ret_260 = vfmlal_high_f16(__s0_260, __s1_260, (float16x4_t) {vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260)}); \
+  __ret_260; \
+})
+#else
+#define vfmlal_lane_high_f16(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \
+  float32x2_t __s0_261 = __p0_261; \
+  float16x4_t __s1_261 = __p1_261; \
+  float16x4_t __s2_261 = __p2_261; \
+  float32x2_t __rev0_261;  __rev0_261 = __builtin_shufflevector(__s0_261, __s0_261, 1, 0); \
+  float16x4_t __rev1_261;  __rev1_261 = __builtin_shufflevector(__s1_261, __s1_261, 3, 2, 1, 0); \
+  float16x4_t __rev2_261;  __rev2_261 = __builtin_shufflevector(__s2_261, __s2_261, 3, 2, 1, 0); \
+  float32x2_t __ret_261; \
+  __ret_261 = __noswap_vfmlal_high_f16(__rev0_261, __rev1_261, (float16x4_t) {__noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261)}); \
+  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 1, 0); \
+  __ret_261; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlalq_lane_low_f16(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \
+  float32x4_t __s0_262 = __p0_262; \
+  float16x8_t __s1_262 = __p1_262; \
+  float16x4_t __s2_262 = __p2_262; \
+  float32x4_t __ret_262; \
+  __ret_262 = vfmlalq_low_f16(__s0_262, __s1_262, (float16x8_t) {vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262)}); \
+  __ret_262; \
+})
+#else
+#define vfmlalq_lane_low_f16(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \
+  float32x4_t __s0_263 = __p0_263; \
+  float16x8_t __s1_263 = __p1_263; \
+  float16x4_t __s2_263 = __p2_263; \
+  float32x4_t __rev0_263;  __rev0_263 = __builtin_shufflevector(__s0_263, __s0_263, 3, 2, 1, 0); \
+  float16x8_t __rev1_263;  __rev1_263 = __builtin_shufflevector(__s1_263, __s1_263, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_263;  __rev2_263 = __builtin_shufflevector(__s2_263, __s2_263, 3, 2, 1, 0); \
+  float32x4_t __ret_263; \
+  __ret_263 = __noswap_vfmlalq_low_f16(__rev0_263, __rev1_263, (float16x8_t) {__noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263)}); \
+  __ret_263 = __builtin_shufflevector(__ret_263, __ret_263, 3, 2, 1, 0); \
+  __ret_263; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlal_lane_low_f16(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
+  float32x2_t __s0_264 = __p0_264; \
+  float16x4_t __s1_264 = __p1_264; \
   float16x4_t __s2_264 = __p2_264; \
-  float32x4_t __ret_264; \
-  __ret_264 = vfmlalq_high_f16(__s0_264, __s1_264, (float16x8_t) {vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264)}); \
+  float32x2_t __ret_264; \
+  __ret_264 = vfmlal_low_f16(__s0_264, __s1_264, (float16x4_t) {vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264)}); \
   __ret_264; \
 })
 #else
-#define vfmlalq_lane_high_f16(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
-  float32x4_t __s0_265 = __p0_265; \
-  float16x8_t __s1_265 = __p1_265; \
+#define vfmlal_lane_low_f16(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
+  float32x2_t __s0_265 = __p0_265; \
+  float16x4_t __s1_265 = __p1_265; \
   float16x4_t __s2_265 = __p2_265; \
-  float32x4_t __rev0_265;  __rev0_265 = __builtin_shufflevector(__s0_265, __s0_265, 3, 2, 1, 0); \
-  float16x8_t __rev1_265;  __rev1_265 = __builtin_shufflevector(__s1_265, __s1_265, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_265;  __rev0_265 = __builtin_shufflevector(__s0_265, __s0_265, 1, 0); \
+  float16x4_t __rev1_265;  __rev1_265 = __builtin_shufflevector(__s1_265, __s1_265, 3, 2, 1, 0); \
   float16x4_t __rev2_265;  __rev2_265 = __builtin_shufflevector(__s2_265, __s2_265, 3, 2, 1, 0); \
-  float32x4_t __ret_265; \
-  __ret_265 = __noswap_vfmlalq_high_f16(__rev0_265, __rev1_265, (float16x8_t) {__noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265)}); \
-  __ret_265 = __builtin_shufflevector(__ret_265, __ret_265, 3, 2, 1, 0); \
+  float32x2_t __ret_265; \
+  __ret_265 = __noswap_vfmlal_low_f16(__rev0_265, __rev1_265, (float16x4_t) {__noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265)}); \
+  __ret_265 = __builtin_shufflevector(__ret_265, __ret_265, 1, 0); \
   __ret_265; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_high_f16(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
-  float32x2_t __s0_266 = __p0_266; \
-  float16x4_t __s1_266 = __p1_266; \
-  float16x4_t __s2_266 = __p2_266; \
-  float32x2_t __ret_266; \
-  __ret_266 = vfmlal_high_f16(__s0_266, __s1_266, (float16x4_t) {vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266)}); \
+#define vfmlalq_laneq_high_f16(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
+  float32x4_t __s0_266 = __p0_266; \
+  float16x8_t __s1_266 = __p1_266; \
+  float16x8_t __s2_266 = __p2_266; \
+  float32x4_t __ret_266; \
+  __ret_266 = vfmlalq_high_f16(__s0_266, __s1_266, (float16x8_t) {vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266)}); \
   __ret_266; \
 })
 #else
-#define vfmlal_lane_high_f16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
-  float32x2_t __s0_267 = __p0_267; \
-  float16x4_t __s1_267 = __p1_267; \
-  float16x4_t __s2_267 = __p2_267; \
-  float32x2_t __rev0_267;  __rev0_267 = __builtin_shufflevector(__s0_267, __s0_267, 1, 0); \
-  float16x4_t __rev1_267;  __rev1_267 = __builtin_shufflevector(__s1_267, __s1_267, 3, 2, 1, 0); \
-  float16x4_t __rev2_267;  __rev2_267 = __builtin_shufflevector(__s2_267, __s2_267, 3, 2, 1, 0); \
-  float32x2_t __ret_267; \
-  __ret_267 = __noswap_vfmlal_high_f16(__rev0_267, __rev1_267, (float16x4_t) {__noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267)}); \
-  __ret_267 = __builtin_shufflevector(__ret_267, __ret_267, 1, 0); \
+#define vfmlalq_laneq_high_f16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
+  float32x4_t __s0_267 = __p0_267; \
+  float16x8_t __s1_267 = __p1_267; \
+  float16x8_t __s2_267 = __p2_267; \
+  float32x4_t __rev0_267;  __rev0_267 = __builtin_shufflevector(__s0_267, __s0_267, 3, 2, 1, 0); \
+  float16x8_t __rev1_267;  __rev1_267 = __builtin_shufflevector(__s1_267, __s1_267, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_267;  __rev2_267 = __builtin_shufflevector(__s2_267, __s2_267, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_267; \
+  __ret_267 = __noswap_vfmlalq_high_f16(__rev0_267, __rev1_267, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267)}); \
+  __ret_267 = __builtin_shufflevector(__ret_267, __ret_267, 3, 2, 1, 0); \
   __ret_267; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_low_f16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
-  float32x4_t __s0_268 = __p0_268; \
-  float16x8_t __s1_268 = __p1_268; \
-  float16x4_t __s2_268 = __p2_268; \
-  float32x4_t __ret_268; \
-  __ret_268 = vfmlalq_low_f16(__s0_268, __s1_268, (float16x8_t) {vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268)}); \
+#define vfmlal_laneq_high_f16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
+  float32x2_t __s0_268 = __p0_268; \
+  float16x4_t __s1_268 = __p1_268; \
+  float16x8_t __s2_268 = __p2_268; \
+  float32x2_t __ret_268; \
+  __ret_268 = vfmlal_high_f16(__s0_268, __s1_268, (float16x4_t) {vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268)}); \
   __ret_268; \
 })
 #else
-#define vfmlalq_lane_low_f16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
-  float32x4_t __s0_269 = __p0_269; \
-  float16x8_t __s1_269 = __p1_269; \
-  float16x4_t __s2_269 = __p2_269; \
-  float32x4_t __rev0_269;  __rev0_269 = __builtin_shufflevector(__s0_269, __s0_269, 3, 2, 1, 0); \
-  float16x8_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_269;  __rev2_269 = __builtin_shufflevector(__s2_269, __s2_269, 3, 2, 1, 0); \
-  float32x4_t __ret_269; \
-  __ret_269 = __noswap_vfmlalq_low_f16(__rev0_269, __rev1_269, (float16x8_t) {__noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269)}); \
-  __ret_269 = __builtin_shufflevector(__ret_269, __ret_269, 3, 2, 1, 0); \
+#define vfmlal_laneq_high_f16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
+  float32x2_t __s0_269 = __p0_269; \
+  float16x4_t __s1_269 = __p1_269; \
+  float16x8_t __s2_269 = __p2_269; \
+  float32x2_t __rev0_269;  __rev0_269 = __builtin_shufflevector(__s0_269, __s0_269, 1, 0); \
+  float16x4_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 3, 2, 1, 0); \
+  float16x8_t __rev2_269;  __rev2_269 = __builtin_shufflevector(__s2_269, __s2_269, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __ret_269; \
+  __ret_269 = __noswap_vfmlal_high_f16(__rev0_269, __rev1_269, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269)}); \
+  __ret_269 = __builtin_shufflevector(__ret_269, __ret_269, 1, 0); \
   __ret_269; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_low_f16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
-  float32x2_t __s0_270 = __p0_270; \
-  float16x4_t __s1_270 = __p1_270; \
-  float16x4_t __s2_270 = __p2_270; \
-  float32x2_t __ret_270; \
-  __ret_270 = vfmlal_low_f16(__s0_270, __s1_270, (float16x4_t) {vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270)}); \
+#define vfmlalq_laneq_low_f16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
+  float32x4_t __s0_270 = __p0_270; \
+  float16x8_t __s1_270 = __p1_270; \
+  float16x8_t __s2_270 = __p2_270; \
+  float32x4_t __ret_270; \
+  __ret_270 = vfmlalq_low_f16(__s0_270, __s1_270, (float16x8_t) {vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270)}); \
   __ret_270; \
 })
 #else
-#define vfmlal_lane_low_f16(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
-  float32x2_t __s0_271 = __p0_271; \
-  float16x4_t __s1_271 = __p1_271; \
-  float16x4_t __s2_271 = __p2_271; \
-  float32x2_t __rev0_271;  __rev0_271 = __builtin_shufflevector(__s0_271, __s0_271, 1, 0); \
-  float16x4_t __rev1_271;  __rev1_271 = __builtin_shufflevector(__s1_271, __s1_271, 3, 2, 1, 0); \
-  float16x4_t __rev2_271;  __rev2_271 = __builtin_shufflevector(__s2_271, __s2_271, 3, 2, 1, 0); \
-  float32x2_t __ret_271; \
-  __ret_271 = __noswap_vfmlal_low_f16(__rev0_271, __rev1_271, (float16x4_t) {__noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271)}); \
-  __ret_271 = __builtin_shufflevector(__ret_271, __ret_271, 1, 0); \
+#define vfmlalq_laneq_low_f16(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
+  float32x4_t __s0_271 = __p0_271; \
+  float16x8_t __s1_271 = __p1_271; \
+  float16x8_t __s2_271 = __p2_271; \
+  float32x4_t __rev0_271;  __rev0_271 = __builtin_shufflevector(__s0_271, __s0_271, 3, 2, 1, 0); \
+  float16x8_t __rev1_271;  __rev1_271 = __builtin_shufflevector(__s1_271, __s1_271, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_271;  __rev2_271 = __builtin_shufflevector(__s2_271, __s2_271, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_271; \
+  __ret_271 = __noswap_vfmlalq_low_f16(__rev0_271, __rev1_271, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271)}); \
+  __ret_271 = __builtin_shufflevector(__ret_271, __ret_271, 3, 2, 1, 0); \
   __ret_271; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_high_f16(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
-  float32x4_t __s0_272 = __p0_272; \
-  float16x8_t __s1_272 = __p1_272; \
+#define vfmlal_laneq_low_f16(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
+  float32x2_t __s0_272 = __p0_272; \
+  float16x4_t __s1_272 = __p1_272; \
   float16x8_t __s2_272 = __p2_272; \
-  float32x4_t __ret_272; \
-  __ret_272 = vfmlalq_high_f16(__s0_272, __s1_272, (float16x8_t) {vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272)}); \
+  float32x2_t __ret_272; \
+  __ret_272 = vfmlal_low_f16(__s0_272, __s1_272, (float16x4_t) {vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272)}); \
   __ret_272; \
 })
 #else
-#define vfmlalq_laneq_high_f16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
-  float32x4_t __s0_273 = __p0_273; \
-  float16x8_t __s1_273 = __p1_273; \
+#define vfmlal_laneq_low_f16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
+  float32x2_t __s0_273 = __p0_273; \
+  float16x4_t __s1_273 = __p1_273; \
   float16x8_t __s2_273 = __p2_273; \
-  float32x4_t __rev0_273;  __rev0_273 = __builtin_shufflevector(__s0_273, __s0_273, 3, 2, 1, 0); \
-  float16x8_t __rev1_273;  __rev1_273 = __builtin_shufflevector(__s1_273, __s1_273, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_273;  __rev0_273 = __builtin_shufflevector(__s0_273, __s0_273, 1, 0); \
+  float16x4_t __rev1_273;  __rev1_273 = __builtin_shufflevector(__s1_273, __s1_273, 3, 2, 1, 0); \
   float16x8_t __rev2_273;  __rev2_273 = __builtin_shufflevector(__s2_273, __s2_273, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_273; \
-  __ret_273 = __noswap_vfmlalq_high_f16(__rev0_273, __rev1_273, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273)}); \
-  __ret_273 = __builtin_shufflevector(__ret_273, __ret_273, 3, 2, 1, 0); \
+  float32x2_t __ret_273; \
+  __ret_273 = __noswap_vfmlal_low_f16(__rev0_273, __rev1_273, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273)}); \
+  __ret_273 = __builtin_shufflevector(__ret_273, __ret_273, 1, 0); \
   __ret_273; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_high_f16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
-  float32x2_t __s0_274 = __p0_274; \
-  float16x4_t __s1_274 = __p1_274; \
-  float16x8_t __s2_274 = __p2_274; \
-  float32x2_t __ret_274; \
-  __ret_274 = vfmlal_high_f16(__s0_274, __s1_274, (float16x4_t) {vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274)}); \
+#define vfmlslq_lane_high_f16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
+  float32x4_t __s0_274 = __p0_274; \
+  float16x8_t __s1_274 = __p1_274; \
+  float16x4_t __s2_274 = __p2_274; \
+  float32x4_t __ret_274; \
+  __ret_274 = vfmlslq_high_f16(__s0_274, __s1_274, (float16x8_t) {vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274)}); \
   __ret_274; \
 })
 #else
-#define vfmlal_laneq_high_f16(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
-  float32x2_t __s0_275 = __p0_275; \
-  float16x4_t __s1_275 = __p1_275; \
-  float16x8_t __s2_275 = __p2_275; \
-  float32x2_t __rev0_275;  __rev0_275 = __builtin_shufflevector(__s0_275, __s0_275, 1, 0); \
-  float16x4_t __rev1_275;  __rev1_275 = __builtin_shufflevector(__s1_275, __s1_275, 3, 2, 1, 0); \
-  float16x8_t __rev2_275;  __rev2_275 = __builtin_shufflevector(__s2_275, __s2_275, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_275; \
-  __ret_275 = __noswap_vfmlal_high_f16(__rev0_275, __rev1_275, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275)}); \
-  __ret_275 = __builtin_shufflevector(__ret_275, __ret_275, 1, 0); \
+#define vfmlslq_lane_high_f16(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
+  float32x4_t __s0_275 = __p0_275; \
+  float16x8_t __s1_275 = __p1_275; \
+  float16x4_t __s2_275 = __p2_275; \
+  float32x4_t __rev0_275;  __rev0_275 = __builtin_shufflevector(__s0_275, __s0_275, 3, 2, 1, 0); \
+  float16x8_t __rev1_275;  __rev1_275 = __builtin_shufflevector(__s1_275, __s1_275, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_275;  __rev2_275 = __builtin_shufflevector(__s2_275, __s2_275, 3, 2, 1, 0); \
+  float32x4_t __ret_275; \
+  __ret_275 = __noswap_vfmlslq_high_f16(__rev0_275, __rev1_275, (float16x8_t) {__noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275)}); \
+  __ret_275 = __builtin_shufflevector(__ret_275, __ret_275, 3, 2, 1, 0); \
   __ret_275; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_low_f16(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
-  float32x4_t __s0_276 = __p0_276; \
-  float16x8_t __s1_276 = __p1_276; \
-  float16x8_t __s2_276 = __p2_276; \
-  float32x4_t __ret_276; \
-  __ret_276 = vfmlalq_low_f16(__s0_276, __s1_276, (float16x8_t) {vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276)}); \
+#define vfmlsl_lane_high_f16(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
+  float32x2_t __s0_276 = __p0_276; \
+  float16x4_t __s1_276 = __p1_276; \
+  float16x4_t __s2_276 = __p2_276; \
+  float32x2_t __ret_276; \
+  __ret_276 = vfmlsl_high_f16(__s0_276, __s1_276, (float16x4_t) {vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276)}); \
   __ret_276; \
 })
 #else
-#define vfmlalq_laneq_low_f16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
-  float32x4_t __s0_277 = __p0_277; \
-  float16x8_t __s1_277 = __p1_277; \
-  float16x8_t __s2_277 = __p2_277; \
-  float32x4_t __rev0_277;  __rev0_277 = __builtin_shufflevector(__s0_277, __s0_277, 3, 2, 1, 0); \
-  float16x8_t __rev1_277;  __rev1_277 = __builtin_shufflevector(__s1_277, __s1_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_277;  __rev2_277 = __builtin_shufflevector(__s2_277, __s2_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_277; \
-  __ret_277 = __noswap_vfmlalq_low_f16(__rev0_277, __rev1_277, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277)}); \
-  __ret_277 = __builtin_shufflevector(__ret_277, __ret_277, 3, 2, 1, 0); \
+#define vfmlsl_lane_high_f16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
+  float32x2_t __s0_277 = __p0_277; \
+  float16x4_t __s1_277 = __p1_277; \
+  float16x4_t __s2_277 = __p2_277; \
+  float32x2_t __rev0_277;  __rev0_277 = __builtin_shufflevector(__s0_277, __s0_277, 1, 0); \
+  float16x4_t __rev1_277;  __rev1_277 = __builtin_shufflevector(__s1_277, __s1_277, 3, 2, 1, 0); \
+  float16x4_t __rev2_277;  __rev2_277 = __builtin_shufflevector(__s2_277, __s2_277, 3, 2, 1, 0); \
+  float32x2_t __ret_277; \
+  __ret_277 = __noswap_vfmlsl_high_f16(__rev0_277, __rev1_277, (float16x4_t) {__noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277)}); \
+  __ret_277 = __builtin_shufflevector(__ret_277, __ret_277, 1, 0); \
   __ret_277; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_low_f16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
-  float32x2_t __s0_278 = __p0_278; \
-  float16x4_t __s1_278 = __p1_278; \
-  float16x8_t __s2_278 = __p2_278; \
-  float32x2_t __ret_278; \
-  __ret_278 = vfmlal_low_f16(__s0_278, __s1_278, (float16x4_t) {vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278)}); \
+#define vfmlslq_lane_low_f16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
+  float32x4_t __s0_278 = __p0_278; \
+  float16x8_t __s1_278 = __p1_278; \
+  float16x4_t __s2_278 = __p2_278; \
+  float32x4_t __ret_278; \
+  __ret_278 = vfmlslq_low_f16(__s0_278, __s1_278, (float16x8_t) {vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278)}); \
   __ret_278; \
 })
 #else
-#define vfmlal_laneq_low_f16(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
-  float32x2_t __s0_279 = __p0_279; \
-  float16x4_t __s1_279 = __p1_279; \
-  float16x8_t __s2_279 = __p2_279; \
-  float32x2_t __rev0_279;  __rev0_279 = __builtin_shufflevector(__s0_279, __s0_279, 1, 0); \
-  float16x4_t __rev1_279;  __rev1_279 = __builtin_shufflevector(__s1_279, __s1_279, 3, 2, 1, 0); \
-  float16x8_t __rev2_279;  __rev2_279 = __builtin_shufflevector(__s2_279, __s2_279, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_279; \
-  __ret_279 = __noswap_vfmlal_low_f16(__rev0_279, __rev1_279, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279)}); \
-  __ret_279 = __builtin_shufflevector(__ret_279, __ret_279, 1, 0); \
+#define vfmlslq_lane_low_f16(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
+  float32x4_t __s0_279 = __p0_279; \
+  float16x8_t __s1_279 = __p1_279; \
+  float16x4_t __s2_279 = __p2_279; \
+  float32x4_t __rev0_279;  __rev0_279 = __builtin_shufflevector(__s0_279, __s0_279, 3, 2, 1, 0); \
+  float16x8_t __rev1_279;  __rev1_279 = __builtin_shufflevector(__s1_279, __s1_279, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_279;  __rev2_279 = __builtin_shufflevector(__s2_279, __s2_279, 3, 2, 1, 0); \
+  float32x4_t __ret_279; \
+  __ret_279 = __noswap_vfmlslq_low_f16(__rev0_279, __rev1_279, (float16x8_t) {__noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279)}); \
+  __ret_279 = __builtin_shufflevector(__ret_279, __ret_279, 3, 2, 1, 0); \
   __ret_279; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_high_f16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
-  float32x4_t __s0_280 = __p0_280; \
-  float16x8_t __s1_280 = __p1_280; \
+#define vfmlsl_lane_low_f16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
+  float32x2_t __s0_280 = __p0_280; \
+  float16x4_t __s1_280 = __p1_280; \
   float16x4_t __s2_280 = __p2_280; \
-  float32x4_t __ret_280; \
-  __ret_280 = vfmlslq_high_f16(__s0_280, __s1_280, (float16x8_t) {vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280)}); \
+  float32x2_t __ret_280; \
+  __ret_280 = vfmlsl_low_f16(__s0_280, __s1_280, (float16x4_t) {vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280)}); \
   __ret_280; \
 })
 #else
-#define vfmlslq_lane_high_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
-  float32x4_t __s0_281 = __p0_281; \
-  float16x8_t __s1_281 = __p1_281; \
+#define vfmlsl_lane_low_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
+  float32x2_t __s0_281 = __p0_281; \
+  float16x4_t __s1_281 = __p1_281; \
   float16x4_t __s2_281 = __p2_281; \
-  float32x4_t __rev0_281;  __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 3, 2, 1, 0); \
-  float16x8_t __rev1_281;  __rev1_281 = __builtin_shufflevector(__s1_281, __s1_281, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_281;  __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 1, 0); \
+  float16x4_t __rev1_281;  __rev1_281 = __builtin_shufflevector(__s1_281, __s1_281, 3, 2, 1, 0); \
   float16x4_t __rev2_281;  __rev2_281 = __builtin_shufflevector(__s2_281, __s2_281, 3, 2, 1, 0); \
-  float32x4_t __ret_281; \
-  __ret_281 = __noswap_vfmlslq_high_f16(__rev0_281, __rev1_281, (float16x8_t) {__noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281)}); \
-  __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 3, 2, 1, 0); \
+  float32x2_t __ret_281; \
+  __ret_281 = __noswap_vfmlsl_low_f16(__rev0_281, __rev1_281, (float16x4_t) {__noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281)}); \
+  __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 1, 0); \
   __ret_281; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_high_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
-  float32x2_t __s0_282 = __p0_282; \
-  float16x4_t __s1_282 = __p1_282; \
-  float16x4_t __s2_282 = __p2_282; \
-  float32x2_t __ret_282; \
-  __ret_282 = vfmlsl_high_f16(__s0_282, __s1_282, (float16x4_t) {vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282)}); \
+#define vfmlslq_laneq_high_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
+  float32x4_t __s0_282 = __p0_282; \
+  float16x8_t __s1_282 = __p1_282; \
+  float16x8_t __s2_282 = __p2_282; \
+  float32x4_t __ret_282; \
+  __ret_282 = vfmlslq_high_f16(__s0_282, __s1_282, (float16x8_t) {vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282)}); \
   __ret_282; \
 })
 #else
-#define vfmlsl_lane_high_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
-  float32x2_t __s0_283 = __p0_283; \
-  float16x4_t __s1_283 = __p1_283; \
-  float16x4_t __s2_283 = __p2_283; \
-  float32x2_t __rev0_283;  __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 1, 0); \
-  float16x4_t __rev1_283;  __rev1_283 = __builtin_shufflevector(__s1_283, __s1_283, 3, 2, 1, 0); \
-  float16x4_t __rev2_283;  __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 3, 2, 1, 0); \
-  float32x2_t __ret_283; \
-  __ret_283 = __noswap_vfmlsl_high_f16(__rev0_283, __rev1_283, (float16x4_t) {__noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283)}); \
-  __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 1, 0); \
+#define vfmlslq_laneq_high_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
+  float32x4_t __s0_283 = __p0_283; \
+  float16x8_t __s1_283 = __p1_283; \
+  float16x8_t __s2_283 = __p2_283; \
+  float32x4_t __rev0_283;  __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 3, 2, 1, 0); \
+  float16x8_t __rev1_283;  __rev1_283 = __builtin_shufflevector(__s1_283, __s1_283, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_283;  __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_283; \
+  __ret_283 = __noswap_vfmlslq_high_f16(__rev0_283, __rev1_283, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283)}); \
+  __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 3, 2, 1, 0); \
   __ret_283; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_low_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
-  float32x4_t __s0_284 = __p0_284; \
-  float16x8_t __s1_284 = __p1_284; \
-  float16x4_t __s2_284 = __p2_284; \
-  float32x4_t __ret_284; \
-  __ret_284 = vfmlslq_low_f16(__s0_284, __s1_284, (float16x8_t) {vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284)}); \
+#define vfmlsl_laneq_high_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
+  float32x2_t __s0_284 = __p0_284; \
+  float16x4_t __s1_284 = __p1_284; \
+  float16x8_t __s2_284 = __p2_284; \
+  float32x2_t __ret_284; \
+  __ret_284 = vfmlsl_high_f16(__s0_284, __s1_284, (float16x4_t) {vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284)}); \
   __ret_284; \
 })
 #else
-#define vfmlslq_lane_low_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
-  float32x4_t __s0_285 = __p0_285; \
-  float16x8_t __s1_285 = __p1_285; \
-  float16x4_t __s2_285 = __p2_285; \
-  float32x4_t __rev0_285;  __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 3, 2, 1, 0); \
-  float16x8_t __rev1_285;  __rev1_285 = __builtin_shufflevector(__s1_285, __s1_285, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_285;  __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 3, 2, 1, 0); \
-  float32x4_t __ret_285; \
-  __ret_285 = __noswap_vfmlslq_low_f16(__rev0_285, __rev1_285, (float16x8_t) {__noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285)}); \
-  __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 3, 2, 1, 0); \
+#define vfmlsl_laneq_high_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
+  float32x2_t __s0_285 = __p0_285; \
+  float16x4_t __s1_285 = __p1_285; \
+  float16x8_t __s2_285 = __p2_285; \
+  float32x2_t __rev0_285;  __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 1, 0); \
+  float16x4_t __rev1_285;  __rev1_285 = __builtin_shufflevector(__s1_285, __s1_285, 3, 2, 1, 0); \
+  float16x8_t __rev2_285;  __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __ret_285; \
+  __ret_285 = __noswap_vfmlsl_high_f16(__rev0_285, __rev1_285, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285)}); \
+  __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 1, 0); \
   __ret_285; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_low_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
-  float32x2_t __s0_286 = __p0_286; \
-  float16x4_t __s1_286 = __p1_286; \
-  float16x4_t __s2_286 = __p2_286; \
-  float32x2_t __ret_286; \
-  __ret_286 = vfmlsl_low_f16(__s0_286, __s1_286, (float16x4_t) {vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286)}); \
+#define vfmlslq_laneq_low_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
+  float32x4_t __s0_286 = __p0_286; \
+  float16x8_t __s1_286 = __p1_286; \
+  float16x8_t __s2_286 = __p2_286; \
+  float32x4_t __ret_286; \
+  __ret_286 = vfmlslq_low_f16(__s0_286, __s1_286, (float16x8_t) {vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286)}); \
   __ret_286; \
 })
 #else
-#define vfmlsl_lane_low_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
-  float32x2_t __s0_287 = __p0_287; \
-  float16x4_t __s1_287 = __p1_287; \
-  float16x4_t __s2_287 = __p2_287; \
-  float32x2_t __rev0_287;  __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 1, 0); \
-  float16x4_t __rev1_287;  __rev1_287 = __builtin_shufflevector(__s1_287, __s1_287, 3, 2, 1, 0); \
-  float16x4_t __rev2_287;  __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 3, 2, 1, 0); \
-  float32x2_t __ret_287; \
-  __ret_287 = __noswap_vfmlsl_low_f16(__rev0_287, __rev1_287, (float16x4_t) {__noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287)}); \
-  __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 1, 0); \
+#define vfmlslq_laneq_low_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
+  float32x4_t __s0_287 = __p0_287; \
+  float16x8_t __s1_287 = __p1_287; \
+  float16x8_t __s2_287 = __p2_287; \
+  float32x4_t __rev0_287;  __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 3, 2, 1, 0); \
+  float16x8_t __rev1_287;  __rev1_287 = __builtin_shufflevector(__s1_287, __s1_287, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_287;  __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_287; \
+  __ret_287 = __noswap_vfmlslq_low_f16(__rev0_287, __rev1_287, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287)}); \
+  __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 3, 2, 1, 0); \
   __ret_287; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_high_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
-  float32x4_t __s0_288 = __p0_288; \
-  float16x8_t __s1_288 = __p1_288; \
+#define vfmlsl_laneq_low_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
+  float32x2_t __s0_288 = __p0_288; \
+  float16x4_t __s1_288 = __p1_288; \
   float16x8_t __s2_288 = __p2_288; \
-  float32x4_t __ret_288; \
-  __ret_288 = vfmlslq_high_f16(__s0_288, __s1_288, (float16x8_t) {vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288)}); \
+  float32x2_t __ret_288; \
+  __ret_288 = vfmlsl_low_f16(__s0_288, __s1_288, (float16x4_t) {vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288)}); \
   __ret_288; \
 })
 #else
-#define vfmlslq_laneq_high_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
-  float32x4_t __s0_289 = __p0_289; \
-  float16x8_t __s1_289 = __p1_289; \
+#define vfmlsl_laneq_low_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
+  float32x2_t __s0_289 = __p0_289; \
+  float16x4_t __s1_289 = __p1_289; \
   float16x8_t __s2_289 = __p2_289; \
-  float32x4_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 3, 2, 1, 0); \
-  float16x8_t __rev1_289;  __rev1_289 = __builtin_shufflevector(__s1_289, __s1_289, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 1, 0); \
+  float16x4_t __rev1_289;  __rev1_289 = __builtin_shufflevector(__s1_289, __s1_289, 3, 2, 1, 0); \
   float16x8_t __rev2_289;  __rev2_289 = __builtin_shufflevector(__s2_289, __s2_289, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_289; \
-  __ret_289 = __noswap_vfmlslq_high_f16(__rev0_289, __rev1_289, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289)}); \
-  __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 3, 2, 1, 0); \
+  float32x2_t __ret_289; \
+  __ret_289 = __noswap_vfmlsl_low_f16(__rev0_289, __rev1_289, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289)}); \
+  __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 1, 0); \
   __ret_289; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_high_f16(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \
-  float32x2_t __s0_290 = __p0_290; \
-  float16x4_t __s1_290 = __p1_290; \
-  float16x8_t __s2_290 = __p2_290; \
-  float32x2_t __ret_290; \
-  __ret_290 = vfmlsl_high_f16(__s0_290, __s1_290, (float16x4_t) {vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290)}); \
-  __ret_290; \
-})
-#else
-#define vfmlsl_laneq_high_f16(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \
-  float32x2_t __s0_291 = __p0_291; \
-  float16x4_t __s1_291 = __p1_291; \
-  float16x8_t __s2_291 = __p2_291; \
-  float32x2_t __rev0_291;  __rev0_291 = __builtin_shufflevector(__s0_291, __s0_291, 1, 0); \
-  float16x4_t __rev1_291;  __rev1_291 = __builtin_shufflevector(__s1_291, __s1_291, 3, 2, 1, 0); \
-  float16x8_t __rev2_291;  __rev2_291 = __builtin_shufflevector(__s2_291, __s2_291, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_291; \
-  __ret_291 = __noswap_vfmlsl_high_f16(__rev0_291, __rev1_291, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291)}); \
-  __ret_291 = __builtin_shufflevector(__ret_291, __ret_291, 1, 0); \
-  __ret_291; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_low_f16(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \
-  float32x4_t __s0_292 = __p0_292; \
-  float16x8_t __s1_292 = __p1_292; \
-  float16x8_t __s2_292 = __p2_292; \
-  float32x4_t __ret_292; \
-  __ret_292 = vfmlslq_low_f16(__s0_292, __s1_292, (float16x8_t) {vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292)}); \
-  __ret_292; \
-})
-#else
-#define vfmlslq_laneq_low_f16(__p0_293, __p1_293, __p2_293, __p3_293) __extension__ ({ \
-  float32x4_t __s0_293 = __p0_293; \
-  float16x8_t __s1_293 = __p1_293; \
-  float16x8_t __s2_293 = __p2_293; \
-  float32x4_t __rev0_293;  __rev0_293 = __builtin_shufflevector(__s0_293, __s0_293, 3, 2, 1, 0); \
-  float16x8_t __rev1_293;  __rev1_293 = __builtin_shufflevector(__s1_293, __s1_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_293;  __rev2_293 = __builtin_shufflevector(__s2_293, __s2_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_293; \
-  __ret_293 = __noswap_vfmlslq_low_f16(__rev0_293, __rev1_293, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293)}); \
-  __ret_293 = __builtin_shufflevector(__ret_293, __ret_293, 3, 2, 1, 0); \
-  __ret_293; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_low_f16(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
-  float32x2_t __s0_294 = __p0_294; \
-  float16x4_t __s1_294 = __p1_294; \
-  float16x8_t __s2_294 = __p2_294; \
-  float32x2_t __ret_294; \
-  __ret_294 = vfmlsl_low_f16(__s0_294, __s1_294, (float16x4_t) {vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294)}); \
-  __ret_294; \
-})
-#else
-#define vfmlsl_laneq_low_f16(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
-  float32x2_t __s0_295 = __p0_295; \
-  float16x4_t __s1_295 = __p1_295; \
-  float16x8_t __s2_295 = __p2_295; \
-  float32x2_t __rev0_295;  __rev0_295 = __builtin_shufflevector(__s0_295, __s0_295, 1, 0); \
-  float16x4_t __rev1_295;  __rev1_295 = __builtin_shufflevector(__s1_295, __s1_295, 3, 2, 1, 0); \
-  float16x8_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_295; \
-  __ret_295 = __noswap_vfmlsl_low_f16(__rev0_295, __rev1_295, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295)}); \
-  __ret_295 = __builtin_shufflevector(__ret_295, __ret_295, 1, 0); \
-  __ret_295; \
-})
-#endif
-
 #endif
 #if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
-#define vmulh_lane_f16(__p0_296, __p1_296, __p2_296) __extension__ ({ \
-  float16_t __s0_296 = __p0_296; \
-  float16x4_t __s1_296 = __p1_296; \
-  float16_t __ret_296; \
-  __ret_296 = __s0_296 * vget_lane_f16(__s1_296, __p2_296); \
-  __ret_296; \
+#define vmulh_lane_f16(__p0_290, __p1_290, __p2_290) __extension__ ({ \
+  float16_t __s0_290 = __p0_290; \
+  float16x4_t __s1_290 = __p1_290; \
+  float16_t __ret_290; \
+  __ret_290 = __s0_290 * vget_lane_f16(__s1_290, __p2_290); \
+  __ret_290; \
 })
 #else
-#define vmulh_lane_f16(__p0_297, __p1_297, __p2_297) __extension__ ({ \
-  float16_t __s0_297 = __p0_297; \
-  float16x4_t __s1_297 = __p1_297; \
-  float16x4_t __rev1_297;  __rev1_297 = __builtin_shufflevector(__s1_297, __s1_297, 3, 2, 1, 0); \
-  float16_t __ret_297; \
-  __ret_297 = __s0_297 * __noswap_vget_lane_f16(__rev1_297, __p2_297); \
-  __ret_297; \
+#define vmulh_lane_f16(__p0_291, __p1_291, __p2_291) __extension__ ({ \
+  float16_t __s0_291 = __p0_291; \
+  float16x4_t __s1_291 = __p1_291; \
+  float16x4_t __rev1_291;  __rev1_291 = __builtin_shufflevector(__s1_291, __s1_291, 3, 2, 1, 0); \
+  float16_t __ret_291; \
+  __ret_291 = __s0_291 * __noswap_vget_lane_f16(__rev1_291, __p2_291); \
+  __ret_291; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulh_laneq_f16(__p0_298, __p1_298, __p2_298) __extension__ ({ \
-  float16_t __s0_298 = __p0_298; \
-  float16x8_t __s1_298 = __p1_298; \
-  float16_t __ret_298; \
-  __ret_298 = __s0_298 * vgetq_lane_f16(__s1_298, __p2_298); \
-  __ret_298; \
+#define vmulh_laneq_f16(__p0_292, __p1_292, __p2_292) __extension__ ({ \
+  float16_t __s0_292 = __p0_292; \
+  float16x8_t __s1_292 = __p1_292; \
+  float16_t __ret_292; \
+  __ret_292 = __s0_292 * vgetq_lane_f16(__s1_292, __p2_292); \
+  __ret_292; \
 })
 #else
-#define vmulh_laneq_f16(__p0_299, __p1_299, __p2_299) __extension__ ({ \
-  float16_t __s0_299 = __p0_299; \
-  float16x8_t __s1_299 = __p1_299; \
-  float16x8_t __rev1_299;  __rev1_299 = __builtin_shufflevector(__s1_299, __s1_299, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_299; \
-  __ret_299 = __s0_299 * __noswap_vgetq_lane_f16(__rev1_299, __p2_299); \
-  __ret_299; \
+#define vmulh_laneq_f16(__p0_293, __p1_293, __p2_293) __extension__ ({ \
+  float16_t __s0_293 = __p0_293; \
+  float16x8_t __s1_293 = __p1_293; \
+  float16x8_t __rev1_293;  __rev1_293 = __builtin_shufflevector(__s1_293, __s1_293, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16_t __ret_293; \
+  __ret_293 = __s0_293 * __noswap_vgetq_lane_f16(__rev1_293, __p2_293); \
+  __ret_293; \
 })
 #endif
 
 #endif
 #if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
   int32_t __ret;
   __ret = vqadds_s32(__p0, vqrdmulhs_s32(__p1, __p2));
   return __ret;
 }
-#else
-__ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqadds_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
   int16_t __ret;
   __ret = vqaddh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahs_lane_s32(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
+  int32_t __s0_294 = __p0_294; \
+  int32_t __s1_294 = __p1_294; \
+  int32x2_t __s2_294 = __p2_294; \
+  int32_t __ret_294; \
+  __ret_294 = vqadds_s32(__s0_294, vqrdmulhs_s32(__s1_294, vget_lane_s32(__s2_294, __p3_294))); \
+  __ret_294; \
+})
 #else
-__ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqaddh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
+#define vqrdmlahs_lane_s32(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
+  int32_t __s0_295 = __p0_295; \
+  int32_t __s1_295 = __p1_295; \
+  int32x2_t __s2_295 = __p2_295; \
+  int32x2_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 1, 0); \
+  int32_t __ret_295; \
+  __ret_295 = vqadds_s32(__s0_295, vqrdmulhs_s32(__s1_295, __noswap_vget_lane_s32(__rev2_295, __p3_295))); \
+  __ret_295; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_lane_s32(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
-  int32_t __s0_300 = __p0_300; \
-  int32_t __s1_300 = __p1_300; \
-  int32x2_t __s2_300 = __p2_300; \
-  int32_t __ret_300; \
-  __ret_300 = vqadds_s32(__s0_300, vqrdmulhs_s32(__s1_300, vget_lane_s32(__s2_300, __p3_300))); \
+#define vqrdmlahh_lane_s16(__p0_296, __p1_296, __p2_296, __p3_296) __extension__ ({ \
+  int16_t __s0_296 = __p0_296; \
+  int16_t __s1_296 = __p1_296; \
+  int16x4_t __s2_296 = __p2_296; \
+  int16_t __ret_296; \
+  __ret_296 = vqaddh_s16(__s0_296, vqrdmulhh_s16(__s1_296, vget_lane_s16(__s2_296, __p3_296))); \
+  __ret_296; \
+})
+#else
+#define vqrdmlahh_lane_s16(__p0_297, __p1_297, __p2_297, __p3_297) __extension__ ({ \
+  int16_t __s0_297 = __p0_297; \
+  int16_t __s1_297 = __p1_297; \
+  int16x4_t __s2_297 = __p2_297; \
+  int16x4_t __rev2_297;  __rev2_297 = __builtin_shufflevector(__s2_297, __s2_297, 3, 2, 1, 0); \
+  int16_t __ret_297; \
+  __ret_297 = vqaddh_s16(__s0_297, vqrdmulhh_s16(__s1_297, __noswap_vget_lane_s16(__rev2_297, __p3_297))); \
+  __ret_297; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahs_laneq_s32(__p0_298, __p1_298, __p2_298, __p3_298) __extension__ ({ \
+  int32_t __s0_298 = __p0_298; \
+  int32_t __s1_298 = __p1_298; \
+  int32x4_t __s2_298 = __p2_298; \
+  int32_t __ret_298; \
+  __ret_298 = vqadds_s32(__s0_298, vqrdmulhs_s32(__s1_298, vgetq_lane_s32(__s2_298, __p3_298))); \
+  __ret_298; \
+})
+#else
+#define vqrdmlahs_laneq_s32(__p0_299, __p1_299, __p2_299, __p3_299) __extension__ ({ \
+  int32_t __s0_299 = __p0_299; \
+  int32_t __s1_299 = __p1_299; \
+  int32x4_t __s2_299 = __p2_299; \
+  int32x4_t __rev2_299;  __rev2_299 = __builtin_shufflevector(__s2_299, __s2_299, 3, 2, 1, 0); \
+  int32_t __ret_299; \
+  __ret_299 = vqadds_s32(__s0_299, vqrdmulhs_s32(__s1_299, __noswap_vgetq_lane_s32(__rev2_299, __p3_299))); \
+  __ret_299; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahh_laneq_s16(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
+  int16_t __s0_300 = __p0_300; \
+  int16_t __s1_300 = __p1_300; \
+  int16x8_t __s2_300 = __p2_300; \
+  int16_t __ret_300; \
+  __ret_300 = vqaddh_s16(__s0_300, vqrdmulhh_s16(__s1_300, vgetq_lane_s16(__s2_300, __p3_300))); \
   __ret_300; \
 })
 #else
-#define vqrdmlahs_lane_s32(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
-  int32_t __s0_301 = __p0_301; \
-  int32_t __s1_301 = __p1_301; \
-  int32x2_t __s2_301 = __p2_301; \
-  int32x2_t __rev2_301;  __rev2_301 = __builtin_shufflevector(__s2_301, __s2_301, 1, 0); \
-  int32_t __ret_301; \
-  __ret_301 = __noswap_vqadds_s32(__s0_301, __noswap_vqrdmulhs_s32(__s1_301, __noswap_vget_lane_s32(__rev2_301, __p3_301))); \
+#define vqrdmlahh_laneq_s16(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
+  int16_t __s0_301 = __p0_301; \
+  int16_t __s1_301 = __p1_301; \
+  int16x8_t __s2_301 = __p2_301; \
+  int16x8_t __rev2_301;  __rev2_301 = __builtin_shufflevector(__s2_301, __s2_301, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_301; \
+  __ret_301 = vqaddh_s16(__s0_301, vqrdmulhh_s16(__s1_301, __noswap_vgetq_lane_s16(__rev2_301, __p3_301))); \
   __ret_301; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_lane_s16(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
-  int16_t __s0_302 = __p0_302; \
-  int16_t __s1_302 = __p1_302; \
-  int16x4_t __s2_302 = __p2_302; \
-  int16_t __ret_302; \
-  __ret_302 = vqaddh_s16(__s0_302, vqrdmulhh_s16(__s1_302, vget_lane_s16(__s2_302, __p3_302))); \
-  __ret_302; \
-})
-#else
-#define vqrdmlahh_lane_s16(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
-  int16_t __s0_303 = __p0_303; \
-  int16_t __s1_303 = __p1_303; \
-  int16x4_t __s2_303 = __p2_303; \
-  int16x4_t __rev2_303;  __rev2_303 = __builtin_shufflevector(__s2_303, __s2_303, 3, 2, 1, 0); \
-  int16_t __ret_303; \
-  __ret_303 = __noswap_vqaddh_s16(__s0_303, __noswap_vqrdmulhh_s16(__s1_303, __noswap_vget_lane_s16(__rev2_303, __p3_303))); \
-  __ret_303; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_laneq_s32(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
-  int32_t __s0_304 = __p0_304; \
-  int32_t __s1_304 = __p1_304; \
-  int32x4_t __s2_304 = __p2_304; \
-  int32_t __ret_304; \
-  __ret_304 = vqadds_s32(__s0_304, vqrdmulhs_s32(__s1_304, vgetq_lane_s32(__s2_304, __p3_304))); \
-  __ret_304; \
-})
-#else
-#define vqrdmlahs_laneq_s32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
-  int32_t __s0_305 = __p0_305; \
-  int32_t __s1_305 = __p1_305; \
-  int32x4_t __s2_305 = __p2_305; \
-  int32x4_t __rev2_305;  __rev2_305 = __builtin_shufflevector(__s2_305, __s2_305, 3, 2, 1, 0); \
-  int32_t __ret_305; \
-  __ret_305 = __noswap_vqadds_s32(__s0_305, __noswap_vqrdmulhs_s32(__s1_305, __noswap_vgetq_lane_s32(__rev2_305, __p3_305))); \
-  __ret_305; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_laneq_s16(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
-  int16_t __s0_306 = __p0_306; \
-  int16_t __s1_306 = __p1_306; \
-  int16x8_t __s2_306 = __p2_306; \
-  int16_t __ret_306; \
-  __ret_306 = vqaddh_s16(__s0_306, vqrdmulhh_s16(__s1_306, vgetq_lane_s16(__s2_306, __p3_306))); \
-  __ret_306; \
-})
-#else
-#define vqrdmlahh_laneq_s16(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
-  int16_t __s0_307 = __p0_307; \
-  int16_t __s1_307 = __p1_307; \
-  int16x8_t __s2_307 = __p2_307; \
-  int16x8_t __rev2_307;  __rev2_307 = __builtin_shufflevector(__s2_307, __s2_307, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_307; \
-  __ret_307 = __noswap_vqaddh_s16(__s0_307, __noswap_vqrdmulhh_s16(__s1_307, __noswap_vgetq_lane_s16(__rev2_307, __p3_307))); \
-  __ret_307; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
   int32_t __ret;
   __ret = vqsubs_s32(__p0, vqrdmulhs_s32(__p1, __p2));
   return __ret;
 }
-#else
-__ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqsubs_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
   int16_t __ret;
   __ret = vqsubh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshs_lane_s32(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
+  int32_t __s0_302 = __p0_302; \
+  int32_t __s1_302 = __p1_302; \
+  int32x2_t __s2_302 = __p2_302; \
+  int32_t __ret_302; \
+  __ret_302 = vqsubs_s32(__s0_302, vqrdmulhs_s32(__s1_302, vget_lane_s32(__s2_302, __p3_302))); \
+  __ret_302; \
+})
 #else
-__ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqsubh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
+#define vqrdmlshs_lane_s32(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
+  int32_t __s0_303 = __p0_303; \
+  int32_t __s1_303 = __p1_303; \
+  int32x2_t __s2_303 = __p2_303; \
+  int32x2_t __rev2_303;  __rev2_303 = __builtin_shufflevector(__s2_303, __s2_303, 1, 0); \
+  int32_t __ret_303; \
+  __ret_303 = vqsubs_s32(__s0_303, vqrdmulhs_s32(__s1_303, __noswap_vget_lane_s32(__rev2_303, __p3_303))); \
+  __ret_303; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_lane_s32(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
-  int32_t __s0_308 = __p0_308; \
-  int32_t __s1_308 = __p1_308; \
-  int32x2_t __s2_308 = __p2_308; \
-  int32_t __ret_308; \
-  __ret_308 = vqsubs_s32(__s0_308, vqrdmulhs_s32(__s1_308, vget_lane_s32(__s2_308, __p3_308))); \
+#define vqrdmlshh_lane_s16(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
+  int16_t __s0_304 = __p0_304; \
+  int16_t __s1_304 = __p1_304; \
+  int16x4_t __s2_304 = __p2_304; \
+  int16_t __ret_304; \
+  __ret_304 = vqsubh_s16(__s0_304, vqrdmulhh_s16(__s1_304, vget_lane_s16(__s2_304, __p3_304))); \
+  __ret_304; \
+})
+#else
+#define vqrdmlshh_lane_s16(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
+  int16_t __s0_305 = __p0_305; \
+  int16_t __s1_305 = __p1_305; \
+  int16x4_t __s2_305 = __p2_305; \
+  int16x4_t __rev2_305;  __rev2_305 = __builtin_shufflevector(__s2_305, __s2_305, 3, 2, 1, 0); \
+  int16_t __ret_305; \
+  __ret_305 = vqsubh_s16(__s0_305, vqrdmulhh_s16(__s1_305, __noswap_vget_lane_s16(__rev2_305, __p3_305))); \
+  __ret_305; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshs_laneq_s32(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
+  int32_t __s0_306 = __p0_306; \
+  int32_t __s1_306 = __p1_306; \
+  int32x4_t __s2_306 = __p2_306; \
+  int32_t __ret_306; \
+  __ret_306 = vqsubs_s32(__s0_306, vqrdmulhs_s32(__s1_306, vgetq_lane_s32(__s2_306, __p3_306))); \
+  __ret_306; \
+})
+#else
+#define vqrdmlshs_laneq_s32(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
+  int32_t __s0_307 = __p0_307; \
+  int32_t __s1_307 = __p1_307; \
+  int32x4_t __s2_307 = __p2_307; \
+  int32x4_t __rev2_307;  __rev2_307 = __builtin_shufflevector(__s2_307, __s2_307, 3, 2, 1, 0); \
+  int32_t __ret_307; \
+  __ret_307 = vqsubs_s32(__s0_307, vqrdmulhs_s32(__s1_307, __noswap_vgetq_lane_s32(__rev2_307, __p3_307))); \
+  __ret_307; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshh_laneq_s16(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
+  int16_t __s0_308 = __p0_308; \
+  int16_t __s1_308 = __p1_308; \
+  int16x8_t __s2_308 = __p2_308; \
+  int16_t __ret_308; \
+  __ret_308 = vqsubh_s16(__s0_308, vqrdmulhh_s16(__s1_308, vgetq_lane_s16(__s2_308, __p3_308))); \
   __ret_308; \
 })
 #else
-#define vqrdmlshs_lane_s32(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
-  int32_t __s0_309 = __p0_309; \
-  int32_t __s1_309 = __p1_309; \
-  int32x2_t __s2_309 = __p2_309; \
-  int32x2_t __rev2_309;  __rev2_309 = __builtin_shufflevector(__s2_309, __s2_309, 1, 0); \
-  int32_t __ret_309; \
-  __ret_309 = __noswap_vqsubs_s32(__s0_309, __noswap_vqrdmulhs_s32(__s1_309, __noswap_vget_lane_s32(__rev2_309, __p3_309))); \
+#define vqrdmlshh_laneq_s16(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
+  int16_t __s0_309 = __p0_309; \
+  int16_t __s1_309 = __p1_309; \
+  int16x8_t __s2_309 = __p2_309; \
+  int16x8_t __rev2_309;  __rev2_309 = __builtin_shufflevector(__s2_309, __s2_309, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_309; \
+  __ret_309 = vqsubh_s16(__s0_309, vqrdmulhh_s16(__s1_309, __noswap_vgetq_lane_s16(__rev2_309, __p3_309))); \
   __ret_309; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_lane_s16(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
-  int16_t __s0_310 = __p0_310; \
-  int16_t __s1_310 = __p1_310; \
-  int16x4_t __s2_310 = __p2_310; \
-  int16_t __ret_310; \
-  __ret_310 = vqsubh_s16(__s0_310, vqrdmulhh_s16(__s1_310, vget_lane_s16(__s2_310, __p3_310))); \
-  __ret_310; \
-})
-#else
-#define vqrdmlshh_lane_s16(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
-  int16_t __s0_311 = __p0_311; \
-  int16_t __s1_311 = __p1_311; \
-  int16x4_t __s2_311 = __p2_311; \
-  int16x4_t __rev2_311;  __rev2_311 = __builtin_shufflevector(__s2_311, __s2_311, 3, 2, 1, 0); \
-  int16_t __ret_311; \
-  __ret_311 = __noswap_vqsubh_s16(__s0_311, __noswap_vqrdmulhh_s16(__s1_311, __noswap_vget_lane_s16(__rev2_311, __p3_311))); \
-  __ret_311; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_laneq_s32(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
-  int32_t __s0_312 = __p0_312; \
-  int32_t __s1_312 = __p1_312; \
-  int32x4_t __s2_312 = __p2_312; \
-  int32_t __ret_312; \
-  __ret_312 = vqsubs_s32(__s0_312, vqrdmulhs_s32(__s1_312, vgetq_lane_s32(__s2_312, __p3_312))); \
-  __ret_312; \
-})
-#else
-#define vqrdmlshs_laneq_s32(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
-  int32_t __s0_313 = __p0_313; \
-  int32_t __s1_313 = __p1_313; \
-  int32x4_t __s2_313 = __p2_313; \
-  int32x4_t __rev2_313;  __rev2_313 = __builtin_shufflevector(__s2_313, __s2_313, 3, 2, 1, 0); \
-  int32_t __ret_313; \
-  __ret_313 = __noswap_vqsubs_s32(__s0_313, __noswap_vqrdmulhs_s32(__s1_313, __noswap_vgetq_lane_s32(__rev2_313, __p3_313))); \
-  __ret_313; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_laneq_s16(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
-  int16_t __s0_314 = __p0_314; \
-  int16_t __s1_314 = __p1_314; \
-  int16x8_t __s2_314 = __p2_314; \
-  int16_t __ret_314; \
-  __ret_314 = vqsubh_s16(__s0_314, vqrdmulhh_s16(__s1_314, vgetq_lane_s16(__s2_314, __p3_314))); \
-  __ret_314; \
-})
-#else
-#define vqrdmlshh_laneq_s16(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
-  int16_t __s0_315 = __p0_315; \
-  int16_t __s1_315 = __p1_315; \
-  int16x8_t __s2_315 = __p2_315; \
-  int16x8_t __rev2_315;  __rev2_315 = __builtin_shufflevector(__s2_315, __s2_315, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_315; \
-  __ret_315 = __noswap_vqsubh_s16(__s0_315, __noswap_vqrdmulhh_s16(__s1_315, __noswap_vgetq_lane_s16(__rev2_315, __p3_315))); \
-  __ret_315; \
-})
-#endif
-
 #endif
 #if defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
@@ -72870,162 +62089,140 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_p64(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
+#define vcopyq_lane_p64(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
+  poly64x2_t __s0_310 = __p0_310; \
+  poly64x1_t __s2_310 = __p2_310; \
+  poly64x2_t __ret_310; \
+  __ret_310 = vsetq_lane_p64(vget_lane_p64(__s2_310, __p3_310), __s0_310, __p1_310); \
+  __ret_310; \
+})
+#else
+#define vcopyq_lane_p64(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
+  poly64x2_t __s0_311 = __p0_311; \
+  poly64x1_t __s2_311 = __p2_311; \
+  poly64x2_t __rev0_311;  __rev0_311 = __builtin_shufflevector(__s0_311, __s0_311, 1, 0); \
+  poly64x2_t __ret_311; \
+  __ret_311 = __noswap_vsetq_lane_p64(vget_lane_p64(__s2_311, __p3_311), __rev0_311, __p1_311); \
+  __ret_311 = __builtin_shufflevector(__ret_311, __ret_311, 1, 0); \
+  __ret_311; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_f64(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
+  float64x2_t __s0_312 = __p0_312; \
+  float64x1_t __s2_312 = __p2_312; \
+  float64x2_t __ret_312; \
+  __ret_312 = vsetq_lane_f64(vget_lane_f64(__s2_312, __p3_312), __s0_312, __p1_312); \
+  __ret_312; \
+})
+#else
+#define vcopyq_lane_f64(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
+  float64x2_t __s0_313 = __p0_313; \
+  float64x1_t __s2_313 = __p2_313; \
+  float64x2_t __rev0_313;  __rev0_313 = __builtin_shufflevector(__s0_313, __s0_313, 1, 0); \
+  float64x2_t __ret_313; \
+  __ret_313 = __noswap_vsetq_lane_f64(vget_lane_f64(__s2_313, __p3_313), __rev0_313, __p1_313); \
+  __ret_313 = __builtin_shufflevector(__ret_313, __ret_313, 1, 0); \
+  __ret_313; \
+})
+#endif
+
+#define vcopy_lane_p64(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
+  poly64x1_t __s0_314 = __p0_314; \
+  poly64x1_t __s2_314 = __p2_314; \
+  poly64x1_t __ret_314; \
+  __ret_314 = vset_lane_p64(vget_lane_p64(__s2_314, __p3_314), __s0_314, __p1_314); \
+  __ret_314; \
+})
+#define vcopy_lane_f64(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
+  float64x1_t __s0_315 = __p0_315; \
+  float64x1_t __s2_315 = __p2_315; \
+  float64x1_t __ret_315; \
+  __ret_315 = vset_lane_f64(vget_lane_f64(__s2_315, __p3_315), __s0_315, __p1_315); \
+  __ret_315; \
+})
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_p64(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
   poly64x2_t __s0_316 = __p0_316; \
-  poly64x1_t __s2_316 = __p2_316; \
+  poly64x2_t __s2_316 = __p2_316; \
   poly64x2_t __ret_316; \
-  __ret_316 = vsetq_lane_p64(vget_lane_p64(__s2_316, __p3_316), __s0_316, __p1_316); \
+  __ret_316 = vsetq_lane_p64(vgetq_lane_p64(__s2_316, __p3_316), __s0_316, __p1_316); \
   __ret_316; \
 })
 #else
-#define vcopyq_lane_p64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
+#define vcopyq_laneq_p64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
   poly64x2_t __s0_317 = __p0_317; \
-  poly64x1_t __s2_317 = __p2_317; \
+  poly64x2_t __s2_317 = __p2_317; \
   poly64x2_t __rev0_317;  __rev0_317 = __builtin_shufflevector(__s0_317, __s0_317, 1, 0); \
+  poly64x2_t __rev2_317;  __rev2_317 = __builtin_shufflevector(__s2_317, __s2_317, 1, 0); \
   poly64x2_t __ret_317; \
-  __ret_317 = __noswap_vsetq_lane_p64(__noswap_vget_lane_p64(__s2_317, __p3_317), __rev0_317, __p1_317); \
+  __ret_317 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_317, __p3_317), __rev0_317, __p1_317); \
   __ret_317 = __builtin_shufflevector(__ret_317, __ret_317, 1, 0); \
   __ret_317; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_f64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
+#define vcopyq_laneq_f64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
   float64x2_t __s0_318 = __p0_318; \
-  float64x1_t __s2_318 = __p2_318; \
+  float64x2_t __s2_318 = __p2_318; \
   float64x2_t __ret_318; \
-  __ret_318 = vsetq_lane_f64(vget_lane_f64(__s2_318, __p3_318), __s0_318, __p1_318); \
+  __ret_318 = vsetq_lane_f64(vgetq_lane_f64(__s2_318, __p3_318), __s0_318, __p1_318); \
   __ret_318; \
 })
 #else
-#define vcopyq_lane_f64(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
+#define vcopyq_laneq_f64(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
   float64x2_t __s0_319 = __p0_319; \
-  float64x1_t __s2_319 = __p2_319; \
+  float64x2_t __s2_319 = __p2_319; \
   float64x2_t __rev0_319;  __rev0_319 = __builtin_shufflevector(__s0_319, __s0_319, 1, 0); \
+  float64x2_t __rev2_319;  __rev2_319 = __builtin_shufflevector(__s2_319, __s2_319, 1, 0); \
   float64x2_t __ret_319; \
-  __ret_319 = __noswap_vsetq_lane_f64(__noswap_vget_lane_f64(__s2_319, __p3_319), __rev0_319, __p1_319); \
+  __ret_319 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_319, __p3_319), __rev0_319, __p1_319); \
   __ret_319 = __builtin_shufflevector(__ret_319, __ret_319, 1, 0); \
   __ret_319; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_p64(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
+#define vcopy_laneq_p64(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
   poly64x1_t __s0_320 = __p0_320; \
-  poly64x1_t __s2_320 = __p2_320; \
+  poly64x2_t __s2_320 = __p2_320; \
   poly64x1_t __ret_320; \
-  __ret_320 = vset_lane_p64(vget_lane_p64(__s2_320, __p3_320), __s0_320, __p1_320); \
+  __ret_320 = vset_lane_p64(vgetq_lane_p64(__s2_320, __p3_320), __s0_320, __p1_320); \
   __ret_320; \
 })
 #else
-#define vcopy_lane_p64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
+#define vcopy_laneq_p64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
   poly64x1_t __s0_321 = __p0_321; \
-  poly64x1_t __s2_321 = __p2_321; \
+  poly64x2_t __s2_321 = __p2_321; \
+  poly64x2_t __rev2_321;  __rev2_321 = __builtin_shufflevector(__s2_321, __s2_321, 1, 0); \
   poly64x1_t __ret_321; \
-  __ret_321 = __noswap_vset_lane_p64(__noswap_vget_lane_p64(__s2_321, __p3_321), __s0_321, __p1_321); \
+  __ret_321 = vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_321, __p3_321), __s0_321, __p1_321); \
   __ret_321; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_f64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
+#define vcopy_laneq_f64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
   float64x1_t __s0_322 = __p0_322; \
-  float64x1_t __s2_322 = __p2_322; \
+  float64x2_t __s2_322 = __p2_322; \
   float64x1_t __ret_322; \
-  __ret_322 = vset_lane_f64(vget_lane_f64(__s2_322, __p3_322), __s0_322, __p1_322); \
+  __ret_322 = vset_lane_f64(vgetq_lane_f64(__s2_322, __p3_322), __s0_322, __p1_322); \
   __ret_322; \
 })
 #else
-#define vcopy_lane_f64(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
+#define vcopy_laneq_f64(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
   float64x1_t __s0_323 = __p0_323; \
-  float64x1_t __s2_323 = __p2_323; \
+  float64x2_t __s2_323 = __p2_323; \
+  float64x2_t __rev2_323;  __rev2_323 = __builtin_shufflevector(__s2_323, __s2_323, 1, 0); \
   float64x1_t __ret_323; \
-  __ret_323 = __noswap_vset_lane_f64(__noswap_vget_lane_f64(__s2_323, __p3_323), __s0_323, __p1_323); \
+  __ret_323 = vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_323, __p3_323), __s0_323, __p1_323); \
   __ret_323; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p64(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \
-  poly64x2_t __s0_324 = __p0_324; \
-  poly64x2_t __s2_324 = __p2_324; \
-  poly64x2_t __ret_324; \
-  __ret_324 = vsetq_lane_p64(vgetq_lane_p64(__s2_324, __p3_324), __s0_324, __p1_324); \
-  __ret_324; \
-})
-#else
-#define vcopyq_laneq_p64(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \
-  poly64x2_t __s0_325 = __p0_325; \
-  poly64x2_t __s2_325 = __p2_325; \
-  poly64x2_t __rev0_325;  __rev0_325 = __builtin_shufflevector(__s0_325, __s0_325, 1, 0); \
-  poly64x2_t __rev2_325;  __rev2_325 = __builtin_shufflevector(__s2_325, __s2_325, 1, 0); \
-  poly64x2_t __ret_325; \
-  __ret_325 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_325, __p3_325), __rev0_325, __p1_325); \
-  __ret_325 = __builtin_shufflevector(__ret_325, __ret_325, 1, 0); \
-  __ret_325; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f64(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \
-  float64x2_t __s0_326 = __p0_326; \
-  float64x2_t __s2_326 = __p2_326; \
-  float64x2_t __ret_326; \
-  __ret_326 = vsetq_lane_f64(vgetq_lane_f64(__s2_326, __p3_326), __s0_326, __p1_326); \
-  __ret_326; \
-})
-#else
-#define vcopyq_laneq_f64(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \
-  float64x2_t __s0_327 = __p0_327; \
-  float64x2_t __s2_327 = __p2_327; \
-  float64x2_t __rev0_327;  __rev0_327 = __builtin_shufflevector(__s0_327, __s0_327, 1, 0); \
-  float64x2_t __rev2_327;  __rev2_327 = __builtin_shufflevector(__s2_327, __s2_327, 1, 0); \
-  float64x2_t __ret_327; \
-  __ret_327 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_327, __p3_327), __rev0_327, __p1_327); \
-  __ret_327 = __builtin_shufflevector(__ret_327, __ret_327, 1, 0); \
-  __ret_327; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p64(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \
-  poly64x1_t __s0_328 = __p0_328; \
-  poly64x2_t __s2_328 = __p2_328; \
-  poly64x1_t __ret_328; \
-  __ret_328 = vset_lane_p64(vgetq_lane_p64(__s2_328, __p3_328), __s0_328, __p1_328); \
-  __ret_328; \
-})
-#else
-#define vcopy_laneq_p64(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \
-  poly64x1_t __s0_329 = __p0_329; \
-  poly64x2_t __s2_329 = __p2_329; \
-  poly64x2_t __rev2_329;  __rev2_329 = __builtin_shufflevector(__s2_329, __s2_329, 1, 0); \
-  poly64x1_t __ret_329; \
-  __ret_329 = __noswap_vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_329, __p3_329), __s0_329, __p1_329); \
-  __ret_329; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f64(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \
-  float64x1_t __s0_330 = __p0_330; \
-  float64x2_t __s2_330 = __p2_330; \
-  float64x1_t __ret_330; \
-  __ret_330 = vset_lane_f64(vgetq_lane_f64(__s2_330, __p3_330), __s0_330, __p1_330); \
-  __ret_330; \
-})
-#else
-#define vcopy_laneq_f64(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \
-  float64x1_t __s0_331 = __p0_331; \
-  float64x2_t __s2_331 = __p2_331; \
-  float64x2_t __rev2_331;  __rev2_331 = __builtin_shufflevector(__s2_331, __s2_331, 1, 0); \
-  float64x1_t __ret_331; \
-  __ret_331 = __noswap_vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_331, __p3_331), __s0_331, __p1_331); \
-  __ret_331; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
   uint16x8_t __ret;
   __ret = vmlal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
@@ -73377,52 +62574,38 @@
 }
 #endif
 
+#define vmulx_lane_f64(__p0_324, __p1_324, __p2_324) __extension__ ({ \
+  float64x1_t __s0_324 = __p0_324; \
+  float64x1_t __s1_324 = __p1_324; \
+  float64x1_t __ret_324; \
+  float64_t __x_324 = vget_lane_f64(__s0_324, 0); \
+  float64_t __y_324 = vget_lane_f64(__s1_324, __p2_324); \
+  float64_t __z_324 = vmulxd_f64(__x_324, __y_324); \
+  __ret_324 = vset_lane_f64(__z_324, __s0_324, __p2_324); \
+  __ret_324; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmulx_lane_f64(__p0_332, __p1_332, __p2_332) __extension__ ({ \
-  float64x1_t __s0_332 = __p0_332; \
-  float64x1_t __s1_332 = __p1_332; \
-  float64x1_t __ret_332; \
-  float64_t __x_332 = vget_lane_f64(__s0_332, 0); \
-  float64_t __y_332 = vget_lane_f64(__s1_332, __p2_332); \
-  float64_t __z_332 = vmulxd_f64(__x_332, __y_332); \
-  __ret_332 = vset_lane_f64(__z_332, __s0_332, __p2_332); \
-  __ret_332; \
+#define vmulx_laneq_f64(__p0_325, __p1_325, __p2_325) __extension__ ({ \
+  float64x1_t __s0_325 = __p0_325; \
+  float64x2_t __s1_325 = __p1_325; \
+  float64x1_t __ret_325; \
+  float64_t __x_325 = vget_lane_f64(__s0_325, 0); \
+  float64_t __y_325 = vgetq_lane_f64(__s1_325, __p2_325); \
+  float64_t __z_325 = vmulxd_f64(__x_325, __y_325); \
+  __ret_325 = vset_lane_f64(__z_325, __s0_325, 0); \
+  __ret_325; \
 })
 #else
-#define vmulx_lane_f64(__p0_333, __p1_333, __p2_333) __extension__ ({ \
-  float64x1_t __s0_333 = __p0_333; \
-  float64x1_t __s1_333 = __p1_333; \
-  float64x1_t __ret_333; \
-  float64_t __x_333 = __noswap_vget_lane_f64(__s0_333, 0); \
-  float64_t __y_333 = __noswap_vget_lane_f64(__s1_333, __p2_333); \
-  float64_t __z_333 = __noswap_vmulxd_f64(__x_333, __y_333); \
-  __ret_333 = __noswap_vset_lane_f64(__z_333, __s0_333, __p2_333); \
-  __ret_333; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_laneq_f64(__p0_334, __p1_334, __p2_334) __extension__ ({ \
-  float64x1_t __s0_334 = __p0_334; \
-  float64x2_t __s1_334 = __p1_334; \
-  float64x1_t __ret_334; \
-  float64_t __x_334 = vget_lane_f64(__s0_334, 0); \
-  float64_t __y_334 = vgetq_lane_f64(__s1_334, __p2_334); \
-  float64_t __z_334 = vmulxd_f64(__x_334, __y_334); \
-  __ret_334 = vset_lane_f64(__z_334, __s0_334, 0); \
-  __ret_334; \
-})
-#else
-#define vmulx_laneq_f64(__p0_335, __p1_335, __p2_335) __extension__ ({ \
-  float64x1_t __s0_335 = __p0_335; \
-  float64x2_t __s1_335 = __p1_335; \
-  float64x2_t __rev1_335;  __rev1_335 = __builtin_shufflevector(__s1_335, __s1_335, 1, 0); \
-  float64x1_t __ret_335; \
-  float64_t __x_335 = __noswap_vget_lane_f64(__s0_335, 0); \
-  float64_t __y_335 = __noswap_vgetq_lane_f64(__rev1_335, __p2_335); \
-  float64_t __z_335 = __noswap_vmulxd_f64(__x_335, __y_335); \
-  __ret_335 = __noswap_vset_lane_f64(__z_335, __s0_335, 0); \
-  __ret_335; \
+#define vmulx_laneq_f64(__p0_326, __p1_326, __p2_326) __extension__ ({ \
+  float64x1_t __s0_326 = __p0_326; \
+  float64x2_t __s1_326 = __p1_326; \
+  float64x2_t __rev1_326;  __rev1_326 = __builtin_shufflevector(__s1_326, __s1_326, 1, 0); \
+  float64x1_t __ret_326; \
+  float64_t __x_326 = vget_lane_f64(__s0_326, 0); \
+  float64_t __y_326 = __noswap_vgetq_lane_f64(__rev1_326, __p2_326); \
+  float64_t __z_326 = vmulxd_f64(__x_326, __y_326); \
+  __ret_326 = vset_lane_f64(__z_326, __s0_326, 0); \
+  __ret_326; \
 })
 #endif
 
diff --git a/darwin-x86/lib64/clang/9.0.7/include/armintr.h b/darwin-x86/lib64/clang/10.0.1/include/armintr.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/armintr.h
rename to darwin-x86/lib64/clang/10.0.1/include/armintr.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx2intrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx2intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx2intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx2intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512bf16intrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512bf16intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512bf16intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512bf16intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512bitalgintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512bitalgintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512bitalgintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512bitalgintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512bwintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512bwintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512bwintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512bwintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512cdintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512cdintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512cdintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512cdintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512dqintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512dqintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512dqintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512dqintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512erintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512erintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512erintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512erintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512fintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512fintrin.h
index 132761f..4e341a1 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/avx512fintrin.h
@@ -8436,7 +8436,7 @@
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_si512 (__m512i * __P, __m512i __A)
+_mm512_stream_si512 (void * __P, __m512i __A)
 {
   typedef __v8di __v8di_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v8di_aligned)__A, (__v8di_aligned*)__P);
@@ -8450,14 +8450,14 @@
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_pd (double *__P, __m512d __A)
+_mm512_stream_pd (void *__P, __m512d __A)
 {
   typedef __v8df __v8df_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v8df_aligned)__A, (__v8df_aligned*)__P);
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_ps (float *__P, __m512 __A)
+_mm512_stream_ps (void *__P, __m512 __A)
 {
   typedef __v16sf __v16sf_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v16sf_aligned)__A, (__v16sf_aligned*)__P);
@@ -9659,6 +9659,23 @@
 }
 #undef _mm512_mask_reduce_operator
 
+/// Moves the least significant 32 bits of a vector of [16 x i32] to a
+///    32-bit signed integer value.
+///
+/// \headerfile <x86intrin.h>
+///
+/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
+///
+/// \param __A
+///    A vector of [16 x i32]. The least significant 32 bits are moved to the
+///    destination.
+/// \returns A 32-bit signed integer containing the moved value.
+static __inline__ int __DEFAULT_FN_ATTRS512
+_mm512_cvtsi512_si32(__m512i __A) {
+  __v16si __b = (__v16si)__A;
+  return __b[0];
+}
+
 #undef __DEFAULT_FN_ATTRS512
 #undef __DEFAULT_FN_ATTRS128
 #undef __DEFAULT_FN_ATTRS
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512ifmaintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512ifmaintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512ifmaintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512ifmaintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512ifmavlintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512ifmavlintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512ifmavlintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512ifmavlintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512pfintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512pfintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512pfintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512pfintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vbmi2intrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vbmi2intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vbmi2intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vbmi2intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vbmiintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vbmiintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vbmiintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vbmiintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vbmivlintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vbmivlintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vbmivlintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vbmivlintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlbf16intrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlbf16intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlbf16intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlbf16intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlbitalgintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlbitalgintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlbitalgintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlbitalgintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlbwintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlbwintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlbwintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlbwintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlcdintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlcdintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlcdintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlcdintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vldqintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vldqintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vldqintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vldqintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlvbmi2intrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlvbmi2intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlvbmi2intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlvbmi2intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlvnniintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlvnniintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlvnniintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlvnniintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vlvp2intersectintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vlvp2intersectintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vlvp2intersectintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vlvp2intersectintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vnniintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vnniintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vnniintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vnniintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vp2intersectintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vp2intersectintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vp2intersectintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vp2intersectintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vpopcntdqintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vpopcntdqintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vpopcntdqintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vpopcntdqintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512vpopcntdqvlintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avx512vpopcntdqvlintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avx512vpopcntdqvlintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avx512vpopcntdqvlintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avxintrin.h b/darwin-x86/lib64/clang/10.0.1/include/avxintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/avxintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/avxintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/bits/stdatomic.h b/darwin-x86/lib64/clang/10.0.1/include/bits/stdatomic.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/bits/stdatomic.h
rename to darwin-x86/lib64/clang/10.0.1/include/bits/stdatomic.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/bmi2intrin.h b/darwin-x86/lib64/clang/10.0.1/include/bmi2intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/bmi2intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/bmi2intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/bmiintrin.h b/darwin-x86/lib64/clang/10.0.1/include/bmiintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/bmiintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/bmiintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cetintrin.h b/darwin-x86/lib64/clang/10.0.1/include/cetintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/cetintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/cetintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cldemoteintrin.h b/darwin-x86/lib64/clang/10.0.1/include/cldemoteintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/cldemoteintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/cldemoteintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/clflushoptintrin.h b/darwin-x86/lib64/clang/10.0.1/include/clflushoptintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/clflushoptintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/clflushoptintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/clwbintrin.h b/darwin-x86/lib64/clang/10.0.1/include/clwbintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/clwbintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/clwbintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/clzerointrin.h b/darwin-x86/lib64/clang/10.0.1/include/clzerointrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/clzerointrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/clzerointrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cpuid.h b/darwin-x86/lib64/clang/10.0.1/include/cpuid.h
similarity index 98%
rename from darwin-x86/lib64/clang/9.0.7/include/cpuid.h
rename to darwin-x86/lib64/clang/10.0.1/include/cpuid.h
index 02ffac2..4ddd648 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/cpuid.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/cpuid.h
@@ -38,8 +38,8 @@
 #define signature_TM2_ecx 0x3638784d
 /* NSC:     "Geode by NSC" */
 #define signature_NSC_ebx 0x646f6547
-#define signature_NSC_edx 0x43534e20
-#define signature_NSC_ecx 0x79622065
+#define signature_NSC_edx 0x79622065
+#define signature_NSC_ecx 0x43534e20
 /* NEXGEN:  "NexGenDriven" */
 #define signature_NEXGEN_ebx 0x4778654e
 #define signature_NEXGEN_edx 0x72446e65
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/algorithm b/darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/algorithm
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/algorithm
rename to darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/algorithm
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/complex b/darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/complex
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/complex
rename to darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/complex
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/new b/darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/new
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/cuda_wrappers/new
rename to darwin-x86/lib64/clang/10.0.1/include/cuda_wrappers/new
diff --git a/darwin-x86/lib64/clang/9.0.7/include/emmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/emmintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/emmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/emmintrin.h
index 3d55f5f..c8fefdf 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/emmintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/emmintrin.h
@@ -4029,7 +4029,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si64(void const *__p, __m128i __b)
+_mm_storeu_si64(void *__p, __m128i __b)
 {
   struct __storeu_si64 {
     long long __v;
@@ -4050,7 +4050,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si32(void const *__p, __m128i __b)
+_mm_storeu_si32(void *__p, __m128i __b)
 {
   struct __storeu_si32 {
     int __v;
@@ -4071,7 +4071,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si16(void const *__p, __m128i __b)
+_mm_storeu_si16(void *__p, __m128i __b)
 {
   struct __storeu_si16 {
     short __v;
diff --git a/darwin-x86/lib64/clang/9.0.7/include/enqcmdintrin.h b/darwin-x86/lib64/clang/10.0.1/include/enqcmdintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/enqcmdintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/enqcmdintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/f16cintrin.h b/darwin-x86/lib64/clang/10.0.1/include/f16cintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/f16cintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/f16cintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/float.h b/darwin-x86/lib64/clang/10.0.1/include/float.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/float.h
rename to darwin-x86/lib64/clang/10.0.1/include/float.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/fma4intrin.h b/darwin-x86/lib64/clang/10.0.1/include/fma4intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/fma4intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/fma4intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/fmaintrin.h b/darwin-x86/lib64/clang/10.0.1/include/fmaintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/fmaintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/fmaintrin.h
diff --git a/darwin-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h b/darwin-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h
new file mode 100644
index 0000000..59ef284
--- /dev/null
+++ b/darwin-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h
@@ -0,0 +1,299 @@
+//===- FuzzedDataProvider.h - Utility header for fuzz targets ---*- C++ -* ===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+// A single header library providing an utility class to break up an array of
+// bytes. Whenever run on the same input, provides the same output, as long as
+// its methods are called in the same order, with the same arguments.
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
+#define LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
+
+#include <algorithm>
+#include <climits>
+#include <cstddef>
+#include <cstdint>
+#include <cstring>
+#include <initializer_list>
+#include <string>
+#include <type_traits>
+#include <utility>
+#include <vector>
+
+// In addition to the comments below, the API is also briefly documented at
+// https://github.com/google/fuzzing/blob/master/docs/split-inputs.md#fuzzed-data-provider
+class FuzzedDataProvider {
+ public:
+  // |data| is an array of length |size| that the FuzzedDataProvider wraps to
+  // provide more granular access. |data| must outlive the FuzzedDataProvider.
+  FuzzedDataProvider(const uint8_t *data, size_t size)
+      : data_ptr_(data), remaining_bytes_(size) {}
+  ~FuzzedDataProvider() = default;
+
+  // Returns a std::vector containing |num_bytes| of input data. If fewer than
+  // |num_bytes| of data remain, returns a shorter std::vector containing all
+  // of the data that's left. Can be used with any byte sized type, such as
+  // char, unsigned char, uint8_t, etc.
+  template <typename T> std::vector<T> ConsumeBytes(size_t num_bytes) {
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    return ConsumeBytes<T>(num_bytes, num_bytes);
+  }
+
+  // Similar to |ConsumeBytes|, but also appends the terminator value at the end
+  // of the resulting vector. Useful, when a mutable null-terminated C-string is
+  // needed, for example. But that is a rare case. Better avoid it, if possible,
+  // and prefer using |ConsumeBytes| or |ConsumeBytesAsString| methods.
+  template <typename T>
+  std::vector<T> ConsumeBytesWithTerminator(size_t num_bytes,
+                                            T terminator = 0) {
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    std::vector<T> result = ConsumeBytes<T>(num_bytes + 1, num_bytes);
+    result.back() = terminator;
+    return result;
+  }
+
+  // Returns a std::string containing |num_bytes| of input data. Using this and
+  // |.c_str()| on the resulting string is the best way to get an immutable
+  // null-terminated C string. If fewer than |num_bytes| of data remain, returns
+  // a shorter std::string containing all of the data that's left.
+  std::string ConsumeBytesAsString(size_t num_bytes) {
+    static_assert(sizeof(std::string::value_type) == sizeof(uint8_t),
+                  "ConsumeBytesAsString cannot convert the data to a string.");
+
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    std::string result(
+        reinterpret_cast<const std::string::value_type *>(data_ptr_),
+        num_bytes);
+    Advance(num_bytes);
+    return result;
+  }
+
+  // Returns a number in the range [min, max] by consuming bytes from the
+  // input data. The value might not be uniformly distributed in the given
+  // range. If there's no input data left, always returns |min|. |min| must
+  // be less than or equal to |max|.
+  template <typename T> T ConsumeIntegralInRange(T min, T max) {
+    static_assert(std::is_integral<T>::value, "An integral type is required.");
+    static_assert(sizeof(T) <= sizeof(uint64_t), "Unsupported integral type.");
+
+    if (min > max)
+      abort();
+
+    // Use the biggest type possible to hold the range and the result.
+    uint64_t range = static_cast<uint64_t>(max) - min;
+    uint64_t result = 0;
+    size_t offset = 0;
+
+    while (offset < sizeof(T) * CHAR_BIT && (range >> offset) > 0 &&
+           remaining_bytes_ != 0) {
+      // Pull bytes off the end of the seed data. Experimentally, this seems to
+      // allow the fuzzer to more easily explore the input space. This makes
+      // sense, since it works by modifying inputs that caused new code to run,
+      // and this data is often used to encode length of data read by
+      // |ConsumeBytes|. Separating out read lengths makes it easier modify the
+      // contents of the data that is actually read.
+      --remaining_bytes_;
+      result = (result << CHAR_BIT) | data_ptr_[remaining_bytes_];
+      offset += CHAR_BIT;
+    }
+
+    // Avoid division by 0, in case |range + 1| results in overflow.
+    if (range != std::numeric_limits<decltype(range)>::max())
+      result = result % (range + 1);
+
+    return static_cast<T>(min + result);
+  }
+
+  // Returns a std::string of length from 0 to |max_length|. When it runs out of
+  // input data, returns what remains of the input. Designed to be more stable
+  // with respect to a fuzzer inserting characters than just picking a random
+  // length and then consuming that many bytes with |ConsumeBytes|.
+  std::string ConsumeRandomLengthString(size_t max_length) {
+    // Reads bytes from the start of |data_ptr_|. Maps "\\" to "\", and maps "\"
+    // followed by anything else to the end of the string. As a result of this
+    // logic, a fuzzer can insert characters into the string, and the string
+    // will be lengthened to include those new characters, resulting in a more
+    // stable fuzzer than picking the length of a string independently from
+    // picking its contents.
+    std::string result;
+
+    // Reserve the anticipated capaticity to prevent several reallocations.
+    result.reserve(std::min(max_length, remaining_bytes_));
+    for (size_t i = 0; i < max_length && remaining_bytes_ != 0; ++i) {
+      char next = ConvertUnsignedToSigned<char>(data_ptr_[0]);
+      Advance(1);
+      if (next == '\\' && remaining_bytes_ != 0) {
+        next = ConvertUnsignedToSigned<char>(data_ptr_[0]);
+        Advance(1);
+        if (next != '\\')
+          break;
+      }
+      result += next;
+    }
+
+    result.shrink_to_fit();
+    return result;
+  }
+
+  // Returns a std::vector containing all remaining bytes of the input data.
+  template <typename T> std::vector<T> ConsumeRemainingBytes() {
+    return ConsumeBytes<T>(remaining_bytes_);
+  }
+
+  // Returns a std::string containing all remaining bytes of the input data.
+  // Prefer using |ConsumeRemainingBytes| unless you actually need a std::string
+  // object.
+  std::string ConsumeRemainingBytesAsString() {
+    return ConsumeBytesAsString(remaining_bytes_);
+  }
+
+  // Returns a number in the range [Type's min, Type's max]. The value might
+  // not be uniformly distributed in the given range. If there's no input data
+  // left, always returns |min|.
+  template <typename T> T ConsumeIntegral() {
+    return ConsumeIntegralInRange(std::numeric_limits<T>::min(),
+                                  std::numeric_limits<T>::max());
+  }
+
+  // Reads one byte and returns a bool, or false when no data remains.
+  bool ConsumeBool() { return 1 & ConsumeIntegral<uint8_t>(); }
+
+  // Returns a copy of the value selected from the given fixed-size |array|.
+  template <typename T, size_t size>
+  T PickValueInArray(const T (&array)[size]) {
+    static_assert(size > 0, "The array must be non empty.");
+    return array[ConsumeIntegralInRange<size_t>(0, size - 1)];
+  }
+
+  template <typename T>
+  T PickValueInArray(std::initializer_list<const T> list) {
+    // TODO(Dor1s): switch to static_assert once C++14 is allowed.
+    if (!list.size())
+      abort();
+
+    return *(list.begin() + ConsumeIntegralInRange<size_t>(0, list.size() - 1));
+  }
+
+  // Returns an enum value. The enum must start at 0 and be contiguous. It must
+  // also contain |kMaxValue| aliased to its largest (inclusive) value. Such as:
+  // enum class Foo { SomeValue, OtherValue, kMaxValue = OtherValue };
+  template <typename T> T ConsumeEnum() {
+    static_assert(std::is_enum<T>::value, "|T| must be an enum type.");
+    return static_cast<T>(ConsumeIntegralInRange<uint32_t>(
+        0, static_cast<uint32_t>(T::kMaxValue)));
+  }
+
+  // Returns a floating point number in the range [0.0, 1.0]. If there's no
+  // input data left, always returns 0.
+  template <typename T> T ConsumeProbability() {
+    static_assert(std::is_floating_point<T>::value,
+                  "A floating point type is required.");
+
+    // Use different integral types for different floating point types in order
+    // to provide better density of the resulting values.
+    using IntegralType =
+        typename std::conditional<sizeof(T) <= sizeof(uint32_t), uint32_t,
+                                  uint64_t>::type;
+
+    T result = static_cast<T>(ConsumeIntegral<IntegralType>());
+    result /= static_cast<T>(std::numeric_limits<IntegralType>::max());
+    return result;
+  }
+
+  // Returns a floating point value in the range [Type's lowest, Type's max] by
+  // consuming bytes from the input data. If there's no input data left, always
+  // returns approximately 0.
+  template <typename T> T ConsumeFloatingPoint() {
+    return ConsumeFloatingPointInRange<T>(std::numeric_limits<T>::lowest(),
+                                          std::numeric_limits<T>::max());
+  }
+
+  // Returns a floating point value in the given range by consuming bytes from
+  // the input data. If there's no input data left, returns |min|. Note that
+  // |min| must be less than or equal to |max|.
+  template <typename T> T ConsumeFloatingPointInRange(T min, T max) {
+    if (min > max)
+      abort();
+
+    T range = .0;
+    T result = min;
+    constexpr T zero(.0);
+    if (max > zero && min < zero && max > min + std::numeric_limits<T>::max()) {
+      // The diff |max - min| would overflow the given floating point type. Use
+      // the half of the diff as the range and consume a bool to decide whether
+      // the result is in the first of the second part of the diff.
+      range = (max / 2.0) - (min / 2.0);
+      if (ConsumeBool()) {
+        result += range;
+      }
+    } else {
+      range = max - min;
+    }
+
+    return result + range * ConsumeProbability<T>();
+  }
+
+  // Reports the remaining bytes available for fuzzed input.
+  size_t remaining_bytes() { return remaining_bytes_; }
+
+ private:
+  FuzzedDataProvider(const FuzzedDataProvider &) = delete;
+  FuzzedDataProvider &operator=(const FuzzedDataProvider &) = delete;
+
+  void Advance(size_t num_bytes) {
+    if (num_bytes > remaining_bytes_)
+      abort();
+
+    data_ptr_ += num_bytes;
+    remaining_bytes_ -= num_bytes;
+  }
+
+  template <typename T>
+  std::vector<T> ConsumeBytes(size_t size, size_t num_bytes_to_consume) {
+    static_assert(sizeof(T) == sizeof(uint8_t), "Incompatible data type.");
+
+    // The point of using the size-based constructor below is to increase the
+    // odds of having a vector object with capacity being equal to the length.
+    // That part is always implementation specific, but at least both libc++ and
+    // libstdc++ allocate the requested number of bytes in that constructor,
+    // which seems to be a natural choice for other implementations as well.
+    // To increase the odds even more, we also call |shrink_to_fit| below.
+    std::vector<T> result(size);
+    std::memcpy(result.data(), data_ptr_, num_bytes_to_consume);
+    Advance(num_bytes_to_consume);
+
+    // Even though |shrink_to_fit| is also implementation specific, we expect it
+    // to provide an additional assurance in case vector's constructor allocated
+    // a buffer which is larger than the actual amount of data we put inside it.
+    result.shrink_to_fit();
+    return result;
+  }
+
+  template <typename TS, typename TU> TS ConvertUnsignedToSigned(TU value) {
+    static_assert(sizeof(TS) == sizeof(TU), "Incompatible data types.");
+    static_assert(!std::numeric_limits<TU>::is_signed,
+                  "Source type must be unsigned.");
+
+    // TODO(Dor1s): change to `if constexpr` once C++17 becomes mainstream.
+    if (std::numeric_limits<TS>::is_modulo)
+      return static_cast<TS>(value);
+
+    // Avoid using implementation-defined unsigned to signer conversions.
+    // To learn more, see https://stackoverflow.com/questions/13150449.
+    if (value <= std::numeric_limits<TS>::max())
+      return static_cast<TS>(value);
+    else {
+      constexpr auto TS_min = std::numeric_limits<TS>::min();
+      return TS_min + static_cast<char>(value - TS_min);
+    }
+  }
+
+  const uint8_t *data_ptr_;
+  size_t remaining_bytes_;
+};
+
+#endif // LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
diff --git a/darwin-x86/lib64/clang/9.0.7/include/fxsrintrin.h b/darwin-x86/lib64/clang/10.0.1/include/fxsrintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/fxsrintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/fxsrintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/gfniintrin.h b/darwin-x86/lib64/clang/10.0.1/include/gfniintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/gfniintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/gfniintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/htmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/htmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/htmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/htmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/htmxlintrin.h b/darwin-x86/lib64/clang/10.0.1/include/htmxlintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/htmxlintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/htmxlintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ia32intrin.h b/darwin-x86/lib64/clang/10.0.1/include/ia32intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/ia32intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ia32intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/immintrin.h b/darwin-x86/lib64/clang/10.0.1/include/immintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/immintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/immintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/intrin.h b/darwin-x86/lib64/clang/10.0.1/include/intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/inttypes.h b/darwin-x86/lib64/clang/10.0.1/include/inttypes.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/inttypes.h
rename to darwin-x86/lib64/clang/10.0.1/include/inttypes.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/invpcidintrin.h b/darwin-x86/lib64/clang/10.0.1/include/invpcidintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/invpcidintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/invpcidintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/iso646.h b/darwin-x86/lib64/clang/10.0.1/include/iso646.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/iso646.h
rename to darwin-x86/lib64/clang/10.0.1/include/iso646.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/limits.h b/darwin-x86/lib64/clang/10.0.1/include/limits.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/limits.h
rename to darwin-x86/lib64/clang/10.0.1/include/limits.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/lwpintrin.h b/darwin-x86/lib64/clang/10.0.1/include/lwpintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/lwpintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/lwpintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/lzcntintrin.h b/darwin-x86/lib64/clang/10.0.1/include/lzcntintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/lzcntintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/lzcntintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/mm3dnow.h b/darwin-x86/lib64/clang/10.0.1/include/mm3dnow.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/mm3dnow.h
rename to darwin-x86/lib64/clang/10.0.1/include/mm3dnow.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/mm_malloc.h b/darwin-x86/lib64/clang/10.0.1/include/mm_malloc.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/mm_malloc.h
rename to darwin-x86/lib64/clang/10.0.1/include/mm_malloc.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/mmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/mmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/mmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/mmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/module.modulemap b/darwin-x86/lib64/clang/10.0.1/include/module.modulemap
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/module.modulemap
rename to darwin-x86/lib64/clang/10.0.1/include/module.modulemap
diff --git a/darwin-x86/lib64/clang/9.0.7/include/movdirintrin.h b/darwin-x86/lib64/clang/10.0.1/include/movdirintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/movdirintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/movdirintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/msa.h b/darwin-x86/lib64/clang/10.0.1/include/msa.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/msa.h
rename to darwin-x86/lib64/clang/10.0.1/include/msa.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/mwaitxintrin.h b/darwin-x86/lib64/clang/10.0.1/include/mwaitxintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/mwaitxintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/mwaitxintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/nmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/nmmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/nmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/nmmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/omp-tools.h b/darwin-x86/lib64/clang/10.0.1/include/omp-tools.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/omp-tools.h
rename to darwin-x86/lib64/clang/10.0.1/include/omp-tools.h
index d04a2b7..190b538 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/omp-tools.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/omp-tools.h
@@ -1,5 +1,5 @@
 /*
- * include/50/omp-tools.h.var
+ * include/omp-tools.h.var
  */
 
 //===----------------------------------------------------------------------===//
@@ -371,7 +371,7 @@
   ompt_frame_cfa            = 0x10,
   ompt_frame_framepointer   = 0x20,
   ompt_frame_stackaddress   = 0x30
-} ompt_frame_flag_t; 
+} ompt_frame_flag_t;
 
 typedef enum ompt_state_t {
   ompt_state_work_serial                      = 0x000,
@@ -720,14 +720,14 @@
   ompt_data_t *parallel_data,
   ompt_data_t *task_data,
   ompt_dispatch_t kind,
-  ompt_data_t instance 
+  ompt_data_t instance
 );
 
 typedef struct ompt_record_dispatch_t {
   ompt_id_t parallel_id;
   ompt_id_t task_id;
   ompt_dispatch_t kind;
-  ompt_data_t instance; 
+  ompt_data_t instance;
 } ompt_record_dispatch_t;
 
 typedef void (*ompt_callback_task_create_t) (
diff --git a/darwin-x86/lib64/clang/9.0.7/include/omp.h b/darwin-x86/lib64/clang/10.0.1/include/omp.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/omp.h
rename to darwin-x86/lib64/clang/10.0.1/include/omp.h
index 7ba1859..9c37698 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/omp.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/omp.h
@@ -1,5 +1,5 @@
 /*
- * include/50/omp.h.var
+ * include/omp.h.var
  */
 
 
@@ -15,6 +15,9 @@
 #ifndef __OMP_H
 #   define __OMP_H
 
+#   include <stdlib.h>
+#   include <stdint.h>
+
 #   define KMP_VERSION_MAJOR    5
 #   define KMP_VERSION_MINOR    0
 #   define KMP_VERSION_BUILD    20140926
@@ -134,8 +137,6 @@
     extern int  __KAI_KMPC_CONVENTION  omp_get_team_num (void);
     extern int  __KAI_KMPC_CONVENTION  omp_get_cancellation (void);
 
-#   include <stdlib.h>
-#   include <stdint.h>
     /* OpenMP 4.5 */
     extern int   __KAI_KMPC_CONVENTION  omp_get_initial_device (void);
     extern void* __KAI_KMPC_CONVENTION  omp_target_alloc(size_t, int);
diff --git a/darwin-x86/lib64/clang/9.0.7/include/omp-tools.h b/darwin-x86/lib64/clang/10.0.1/include/ompt.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/omp-tools.h
copy to darwin-x86/lib64/clang/10.0.1/include/ompt.h
index d04a2b7..190b538 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/omp-tools.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/ompt.h
@@ -1,5 +1,5 @@
 /*
- * include/50/omp-tools.h.var
+ * include/omp-tools.h.var
  */
 
 //===----------------------------------------------------------------------===//
@@ -371,7 +371,7 @@
   ompt_frame_cfa            = 0x10,
   ompt_frame_framepointer   = 0x20,
   ompt_frame_stackaddress   = 0x30
-} ompt_frame_flag_t; 
+} ompt_frame_flag_t;
 
 typedef enum ompt_state_t {
   ompt_state_work_serial                      = 0x000,
@@ -720,14 +720,14 @@
   ompt_data_t *parallel_data,
   ompt_data_t *task_data,
   ompt_dispatch_t kind,
-  ompt_data_t instance 
+  ompt_data_t instance
 );
 
 typedef struct ompt_record_dispatch_t {
   ompt_id_t parallel_id;
   ompt_id_t task_id;
   ompt_dispatch_t kind;
-  ompt_data_t instance; 
+  ompt_data_t instance;
 } ompt_record_dispatch_t;
 
 typedef void (*ompt_callback_task_create_t) (
diff --git a/darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h b/darwin-x86/lib64/clang/10.0.1/include/opencl-c-base.h
similarity index 97%
rename from darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h
rename to darwin-x86/lib64/clang/10.0.1/include/opencl-c-base.h
index a82954d..430e07d 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/opencl-c-base.h
@@ -126,7 +126,7 @@
 typedef double double16 __attribute__((ext_vector_type(16)));
 #endif
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #define NULL ((void*)0)
 #endif
 
@@ -276,7 +276,7 @@
  */
 #define CLK_GLOBAL_MEM_FENCE   0x02
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 typedef enum memory_scope {
   memory_scope_work_item = __OPENCL_MEMORY_SCOPE_WORK_ITEM,
@@ -288,9 +288,6 @@
 #endif
 } memory_scope;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
 /**
  * Queue a memory fence to ensure correct ordering of memory
  * operations between work-items of a work-group to
@@ -313,7 +310,7 @@
   memory_order_seq_cst = __ATOMIC_SEQ_CST
 } memory_order;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.3, v1.2 s6.12.14, v2.0 s6.13.14 - Image Read and Write Functions
 
@@ -389,14 +386,10 @@
 #endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
 
 // OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #define CLK_NULL_RESERVE_ID (__builtin_astype(((void*)(__SIZE_MAX__)), reserve_id_t))
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
 
 // OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
 #define CL_COMPLETE                                 0x0
 #define CL_RUNNING                                  0x1
 #define CL_SUBMITTED                                0x2
@@ -413,7 +406,7 @@
 #define CLK_OUT_OF_RESOURCES                        -5
 
 #define CLK_NULL_QUEUE                              0
-#define CLK_NULL_EVENT (__builtin_astype(((void*)(__SIZE_MAX__)), clk_event_t))
+#define CLK_NULL_EVENT (__builtin_astype(((__SIZE_MAX__)), clk_event_t))
 
 // execution model related definitions
 #define CLK_ENQUEUE_FLAGS_NO_WAIT                   0x0
@@ -435,7 +428,7 @@
   size_t localWorkSize[MAX_WORK_DIM];
 } ndrange_t;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 #ifdef cl_intel_device_side_avc_motion_estimation
 #pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : begin
diff --git a/darwin-x86/lib64/clang/9.0.7/include/opencl-c.h b/darwin-x86/lib64/clang/10.0.1/include/opencl-c.h
similarity index 98%
rename from darwin-x86/lib64/clang/9.0.7/include/opencl-c.h
rename to darwin-x86/lib64/clang/10.0.1/include/opencl-c.h
index 4207c53..06c5ab6 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/opencl-c.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/opencl-c.h
@@ -11,11 +11,11 @@
 
 #include "opencl-c-base.h"
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifndef cl_khr_depth_images
 #define cl_khr_depth_images
 #endif //cl_khr_depth_images
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 #if __OPENCL_C_VERSION__ < CL_VERSION_2_0
 #ifdef cl_khr_3d_image_writes
@@ -23,10 +23,10 @@
 #endif //cl_khr_3d_image_writes
 #endif //__OPENCL_C_VERSION__ < CL_VERSION_2_0
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 #pragma OPENCL EXTENSION cl_intel_planar_yuv : begin
 #pragma OPENCL EXTENSION cl_intel_planar_yuv : end
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 #define __ovld __attribute__((overloadable))
 #define __conv __attribute__((convergent))
@@ -6517,11 +6517,11 @@
  */
 size_t __ovld __cnfn get_global_offset(uint dimindx);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 size_t __ovld get_enqueued_local_size(uint dimindx);
 size_t __ovld get_global_linear_id(void);
 size_t __ovld get_local_linear_id(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.2, v1.2 s6.12.2, v2.0 s6.13.2 - Math functions
 
@@ -7352,7 +7352,7 @@
  * Returns fmin(x - floor (x), 0x1.fffffep-1f ).
  * floor(x) is returned in iptr.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld fract(float x, float *iptr);
 float2 __ovld fract(float2 x, float2 *iptr);
 float3 __ovld fract(float3 x, float3 *iptr);
@@ -7434,7 +7434,7 @@
 half8 __ovld fract(half8 x, __private half8 *iptr);
 half16 __ovld fract(half16 x, __private half16 *iptr);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Extract mantissa and exponent from x. For each
@@ -7442,7 +7442,7 @@
  * magnitude in the interval [1/2, 1) or 0. Each
  * component of x equals mantissa returned * 2^exp.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld frexp(float x, int *exp);
 float2 __ovld frexp(float2 x, int2 *exp);
 float3 __ovld frexp(float3 x, int3 *exp);
@@ -7524,7 +7524,7 @@
 half8 __ovld frexp(half8 x, __private int8 *exp);
 half16 __ovld frexp(half16 x, __private int16 *exp);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute the value of the square root of x^2 + y^2
@@ -7649,7 +7649,7 @@
 half16 __ovld __cnfn lgamma(half16 x);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld lgamma_r(float x, int *signp);
 float2 __ovld lgamma_r(float2 x, int2 *signp);
 float3 __ovld lgamma_r(float3 x, int3 *signp);
@@ -7731,7 +7731,7 @@
 half8 __ovld lgamma_r(half8 x, __private int8 *signp);
 half16 __ovld lgamma_r(half16 x, __private int16 *signp);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute natural logarithm.
@@ -7955,7 +7955,7 @@
  * the argument. It stores the integral part in the object
  * pointed to by iptr.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld modf(float x, float *iptr);
 float2 __ovld modf(float2 x, float2 *iptr);
 float3 __ovld modf(float3 x, float3 *iptr);
@@ -8037,7 +8037,7 @@
 half8 __ovld modf(half8 x, __private half8 *iptr);
 half16 __ovld modf(half16 x, __private half16 *iptr);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Returns a quiet NaN. The nancode may be placed
@@ -8215,7 +8215,7 @@
  * sign as x/y. It stores this signed value in the object
  * pointed to by quo.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld remquo(float x, float y, int *quo);
 float2 __ovld remquo(float2 x, float2 y, int2 *quo);
 float3 __ovld remquo(float3 x, float3 y, int3 *quo);
@@ -8298,7 +8298,7 @@
 half8 __ovld remquo(half8 x, half8 y, __private int8 *quo);
 half16 __ovld remquo(half16 x, half16 y, __private int16 *quo);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 /**
  * Round to integral value (using round to nearest
  * even rounding mode) in floating-point format.
@@ -8439,7 +8439,7 @@
  * is the return value and computed cosine is returned
  * in cosval.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld sincos(float x, float *cosval);
 float2 __ovld sincos(float2 x, float2 *cosval);
 float3 __ovld sincos(float3 x, float3 *cosval);
@@ -8521,7 +8521,7 @@
 half8 __ovld sincos(half8 x, __private half8 *cosval);
 half16 __ovld sincos(half16 x, __private half16 *cosval);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute hyperbolic sine.
@@ -9446,7 +9446,7 @@
  * returns the size in bits of the type of x or
  * component type of x, if x is a vector.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 char __ovld ctz(char x);
 uchar __ovld ctz(uchar x);
 char2 __ovld ctz(char2 x);
@@ -9495,7 +9495,7 @@
 ulong8 __ovld ctz(ulong8 x);
 long16 __ovld ctz(long16 x);
 ulong16 __ovld ctz(ulong16 x);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Returns mul_hi(a, b) + c.
@@ -11340,7 +11340,7 @@
 half16 __ovld vload16(size_t offset, const __constant half *p);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 char2 __ovld vload2(size_t offset, const char *p);
 uchar2 __ovld vload2(size_t offset, const uchar *p);
 short2 __ovld vload2(size_t offset, const short *p);
@@ -11578,9 +11578,9 @@
 half8 __ovld vload8(size_t offset, const __private half *p);
 half16 __ovld vload16(size_t offset, const __private half *p);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore2(char2 data, size_t offset, char *p);
 void __ovld vstore2(uchar2 data, size_t offset, uchar *p);
 void __ovld vstore2(short2 data, size_t offset, short *p);
@@ -11814,7 +11814,7 @@
 void __ovld vstore8(half8 data, size_t offset, __private half *p);
 void __ovld vstore16(half16 data, size_t offset, __private half *p);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Read sizeof (half) bytes of data from address
@@ -11825,13 +11825,13 @@
  * must be 16-bit aligned.
  */
 float __ovld vload_half(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld vload_half(size_t offset, const half *p);
 #else
 float __ovld vload_half(size_t offset, const __global half *p);
 float __ovld vload_half(size_t offset, const __local half *p);
 float __ovld vload_half(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Read sizeof (halfn) bytes of data from address
@@ -11846,7 +11846,7 @@
 float4 __ovld vload_half4(size_t offset, const __constant half *p);
 float8 __ovld vload_half8(size_t offset, const __constant half *p);
 float16 __ovld vload_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float2 __ovld vload_half2(size_t offset, const half *p);
 float3 __ovld vload_half3(size_t offset, const half *p);
 float4 __ovld vload_half4(size_t offset, const half *p);
@@ -11868,7 +11868,7 @@
 float4 __ovld vload_half4(size_t offset, const __private half *p);
 float8 __ovld vload_half8(size_t offset, const __private half *p);
 float16 __ovld vload_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The float value given by data is first
@@ -11881,7 +11881,7 @@
  * The default current rounding mode is round to
  * nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore_half(float data, size_t offset, half *p);
 void __ovld vstore_half_rte(float data, size_t offset, half *p);
 void __ovld vstore_half_rtz(float data, size_t offset, half *p);
@@ -11927,7 +11927,7 @@
 void __ovld vstore_half_rtp(double data, size_t offset, __private half *p);
 void __ovld vstore_half_rtn(double data, size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The floatn value given by data is converted to
@@ -11940,7 +11940,7 @@
  * The default current rounding mode is round to
  * nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore_half2(float2 data, size_t offset, half *p);
 void __ovld vstore_half3(float3 data, size_t offset, half *p);
 void __ovld vstore_half4(float4 data, size_t offset, half *p);
@@ -12146,7 +12146,7 @@
 void __ovld vstore_half8_rtn(double8 data, size_t offset, __private half *p);
 void __ovld vstore_half16_rtn(double16 data, size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * For n = 1, 2, 4, 8 and 16 read sizeof (halfn)
@@ -12167,7 +12167,7 @@
 float4 __ovld vloada_half4(size_t offset, const __constant half *p);
 float8 __ovld vloada_half8(size_t offset, const __constant half *p);
 float16 __ovld vloada_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld vloada_half(size_t offset, const half *p);
 float2 __ovld vloada_half2(size_t offset, const half *p);
 float3 __ovld vloada_half3(size_t offset, const half *p);
@@ -12193,7 +12193,7 @@
 float4 __ovld vloada_half4(size_t offset, const __private half *p);
 float8 __ovld vloada_half8(size_t offset, const __private half *p);
 float16 __ovld vloada_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The floatn value given by data is converted to
@@ -12211,7 +12211,7 @@
  * mode. The default current rounding mode is
  * round to nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstorea_half(float data, size_t offset, half *p);
 void __ovld vstorea_half2(float2 data, size_t offset, half *p);
 void __ovld vstorea_half3(float3 data, size_t offset, half *p);
@@ -12496,7 +12496,7 @@
 void __ovld vstorea_half8_rtn(double8 data,size_t offset, __private half *p);
 void __ovld vstorea_half16_rtn(double16 data,size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.8, v1.2 s6.12.8, v2.0 s6.13.8 - Synchronization Functions
 
@@ -12532,10 +12532,10 @@
 
 void __ovld __conv barrier(cl_mem_fence_flags flags);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld __conv work_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
 void __ovld __conv work_group_barrier(cl_mem_fence_flags flags);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.9, v1.2 s6.12.9 - Explicit Memory Fence Functions
 
@@ -12580,7 +12580,7 @@
 
 // OpenCL v2.0 s6.13.9 - Address Space Qualifier Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 cl_mem_fence_flags __ovld get_fence(const void *ptr);
 cl_mem_fence_flags __ovld get_fence(void *ptr);
 
@@ -12591,7 +12591,7 @@
  * where gentype is builtin type or user defined type.
  */
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.10, v1.2 s6.12.10, v2.0 s6.13.10 - Async Copies from Global to Local Memory, Local to Global Memory, and Prefetch
 
@@ -13371,7 +13371,7 @@
 
 // OpenCL v2.0 s6.13.11 - Atomics Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // double atomics support requires extensions cl_khr_int64_base_atomics and cl_khr_int64_extended_atomics
 #if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
@@ -13692,7 +13692,7 @@
 void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order);
 void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order, memory_scope scope);
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.12, v1.2 s6.12.12, v2.0 s6.13.12 - Miscellaneous Vector Functions
 
@@ -14186,7 +14186,7 @@
 half16 __ovld __cnfn shuffle2(half16 x, half16 y, ushort16 mask);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 // OpenCL v1.2 s6.12.13, v2.0 s6.13.13 - printf
 
 int printf(__constant const char* st, ...) __attribute__((format(printf, 1, 2)));
@@ -14307,7 +14307,7 @@
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
 float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
 
@@ -14315,7 +14315,7 @@
 int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
 uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, int coord);
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord);
@@ -14325,7 +14325,7 @@
 uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, int coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, float coord);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
 float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
 
@@ -14333,7 +14333,7 @@
 int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 #ifdef cl_khr_depth_images
 float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, float2 coord);
@@ -14358,7 +14358,7 @@
 #endif //cl_khr_gl_msaa_sharing
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord, float lod);
@@ -14410,9 +14410,9 @@
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 /**
 * Sampler-less Image Access
@@ -14447,7 +14447,7 @@
 int4 __purefn __ovld read_imagei(read_only image3d_t image, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, int4 coord);
 
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 // Image read functions returning half4 type
 #ifdef cl_khr_fp16
@@ -14457,7 +14457,7 @@
 half4 __purefn __ovld read_imageh(read_only image2d_t image, sampler_t sampler, float2 coord);
 half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, int4 coord);
 half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, float4 coord);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, int2 coord);
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, float2 coord);
 half4 __purefn __ovld read_imageh(read_only image2d_array_t image, sampler_t sampler, int4 coord);
@@ -14471,11 +14471,11 @@
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, int2 coord);
 half4 __purefn __ovld read_imageh(read_only image2d_array_t image, int4 coord);
 half4 __purefn __ovld read_imageh(read_only image1d_buffer_t image, int coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 #endif //cl_khr_fp16
 
 // Image read functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float4 __purefn __ovld read_imagef(read_write image1d_t image, int coord);
 int4 __purefn __ovld read_imagei(read_write image1d_t image, int coord);
 uint4 __purefn __ovld read_imageui(read_write image1d_t image, int coord);
@@ -14518,7 +14518,7 @@
 float __purefn __ovld read_imagef(read_write image2d_array_msaa_depth_t image, int4 coord, int sample);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 float4 __purefn __ovld read_imagef(read_write image1d_t image, sampler_t sampler, float coord, float lod);
 int4 __purefn __ovld read_imagei(read_write image1d_t image, sampler_t sampler, float coord, float lod);
@@ -14569,7 +14569,7 @@
 uint4 __purefn __ovld read_imageui(read_write image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image read functions returning half4 type
 #ifdef cl_khr_fp16
@@ -14580,7 +14580,7 @@
 half4 __purefn __ovld read_imageh(read_write image2d_array_t image, int4 coord);
 half4 __purefn __ovld read_imageh(read_write image1d_buffer_t image, int coord);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Write color value to location specified by coordinate
@@ -14681,7 +14681,7 @@
 #endif //cl_khr_depth_images
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 void __ovld write_imagef(write_only image1d_t image, int coord, int lod, float4 color);
 void __ovld write_imagei(write_only image1d_t image, int coord, int lod, int4 color);
@@ -14708,7 +14708,7 @@
 void __ovld write_imageui(write_only image3d_t image, int4 coord, int lod, uint4 color);
 #endif
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image write functions for half4 type
 #ifdef cl_khr_fp16
@@ -14723,7 +14723,7 @@
 #endif //cl_khr_fp16
 
 // Image write functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld write_imagef(read_write image2d_t image, int2 coord, float4 color);
 void __ovld write_imagei(read_write image2d_t image, int2 coord, int4 color);
 void __ovld write_imageui(read_write image2d_t image, int2 coord, uint4 color);
@@ -14755,7 +14755,7 @@
 void __ovld write_imagef(read_write image2d_array_depth_t image, int4 coord, float color);
 #endif //cl_khr_depth_images
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 void __ovld write_imagef(read_write image1d_t image, int coord, int lod, float4 color);
 void __ovld write_imagei(read_write image1d_t image, int coord, int lod, int4 color);
@@ -14782,7 +14782,7 @@
 void __ovld write_imageui(read_write image3d_t image, int4 coord, int lod, uint4 color);
 #endif
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image write functions for half4 type
 #ifdef cl_khr_fp16
@@ -14795,7 +14795,7 @@
 void __ovld write_imageh(read_write image2d_array_t image, int4 coord, half4 color);
 void __ovld write_imageh(read_write image1d_buffer_t image, int coord, half4 color);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Note: In OpenCL v1.0/1.1/1.2, image argument of image query builtin functions does not have
 // access qualifier, which by default assume read_only access qualifier. Image query builtin
@@ -14843,7 +14843,7 @@
 int __ovld __cnfn get_image_width(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_width(read_write image1d_t image);
 int __ovld __cnfn get_image_width(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_width(read_write image2d_t image);
@@ -14860,7 +14860,7 @@
 int __ovld __cnfn get_image_width(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_width(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image height in pixels.
@@ -14895,7 +14895,7 @@
 int __ovld __cnfn get_image_height(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_height(read_write image2d_t image);
 int __ovld __cnfn get_image_height(read_write image3d_t image);
 int __ovld __cnfn get_image_height(read_write image2d_array_t image);
@@ -14909,7 +14909,7 @@
 int __ovld __cnfn get_image_height(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_height(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image depth in pixels.
@@ -14920,12 +14920,12 @@
 int __ovld __cnfn get_image_depth(write_only image3d_t image);
 #endif
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_depth(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 /**
  * Return the image miplevels.
@@ -14961,7 +14961,7 @@
 int __ovld get_image_num_mip_levels(read_write image2d_depth_t image);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the channel data type. Valid values are:
@@ -15018,7 +15018,7 @@
 int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_channel_data_type(read_write image1d_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_t image);
@@ -15035,7 +15035,7 @@
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image channel order. Valid values are:
@@ -15090,7 +15090,7 @@
 int __ovld __cnfn get_image_channel_order(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_channel_order(read_write image1d_t image);
 int __ovld __cnfn get_image_channel_order(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_channel_order(read_write image2d_t image);
@@ -15107,7 +15107,7 @@
 int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the 2D image width and height as an int2
@@ -15140,7 +15140,7 @@
 int2 __ovld __cnfn get_image_dim(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int2 __ovld __cnfn get_image_dim(read_write image2d_t image);
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_t image);
 #ifdef cl_khr_depth_images
@@ -15153,7 +15153,7 @@
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_t image);
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the 3D image width, height, and depth as an
@@ -15165,9 +15165,9 @@
 #ifdef cl_khr_3d_image_writes
 int4 __ovld __cnfn get_image_dim(write_only image3d_t image);
 #endif
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int4 __ovld __cnfn get_image_dim(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image array size.
@@ -15193,7 +15193,7 @@
 size_t __ovld __cnfn get_image_array_size(write_only image2d_array_msaa_depth_t image_array);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 size_t __ovld __cnfn get_image_array_size(read_write image1d_array_t image_array);
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_t image_array);
 #ifdef cl_khr_depth_images
@@ -15203,7 +15203,7 @@
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_t image_array);
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_depth_t image_array);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
 * Return the number of samples associated with image
@@ -15219,17 +15219,17 @@
 int __ovld get_image_num_samples(write_only image2d_array_msaa_t image);
 int __ovld get_image_num_samples(write_only image2d_array_msaa_depth_t image);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld get_image_num_samples(read_write image2d_msaa_t image);
 int __ovld get_image_num_samples(read_write image2d_msaa_depth_t image);
 int __ovld get_image_num_samples(read_write image2d_array_msaa_t image);
 int __ovld get_image_num_samples(read_write image2d_array_msaa_depth_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #endif
 
 // OpenCL v2.0 s6.13.15 - Work-group Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __conv work_group_all(int predicate);
 int __ovld __conv work_group_any(int predicate);
 
@@ -15327,16 +15327,16 @@
 double __ovld __conv work_group_scan_inclusive_max(double x);
 #endif //cl_khr_fp64
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 bool __ovld is_valid_reserve_id(reserve_id_t reserve_id);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 
 // OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 ndrange_t __ovld ndrange_1D(size_t);
 ndrange_t __ovld ndrange_1D(size_t, size_t);
@@ -15350,7 +15350,7 @@
 ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3]);
 ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3], const size_t[3]);
 
-int __ovld enqueue_marker(queue_t, uint, const __private clk_event_t*, __private clk_event_t*);
+int __ovld enqueue_marker(queue_t, uint, const clk_event_t*, clk_event_t*);
 
 void __ovld retain_event(clk_event_t);
 
@@ -15365,7 +15365,7 @@
 void __ovld capture_event_profiling_info(clk_event_t, clk_profiling_info, __global void* value);
 
 queue_t __ovld get_default_queue(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL Extension v2.0 s9.17 - Sub-groups
 
@@ -15374,16 +15374,16 @@
 uint    __ovld get_sub_group_size(void);
 uint    __ovld get_max_sub_group_size(void);
 uint    __ovld get_num_sub_groups(void);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld get_enqueued_num_sub_groups(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld get_sub_group_id(void);
 uint    __ovld get_sub_group_local_id(void);
 
 void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 int     __ovld __conv sub_group_all(int predicate);
 int     __ovld __conv sub_group_any(int predicate);
@@ -15573,12 +15573,12 @@
 uint4   __ovld __conv intel_sub_group_block_read4( read_only image2d_t image, int2 coord );
 uint8   __ovld __conv intel_sub_group_block_read8( read_only image2d_t image, int2 coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld __conv intel_sub_group_block_read(read_write image2d_t image, int2 coord);
 uint2   __ovld __conv intel_sub_group_block_read2(read_write image2d_t image, int2 coord);
 uint4   __ovld __conv intel_sub_group_block_read4(read_write image2d_t image, int2 coord);
 uint8   __ovld __conv intel_sub_group_block_read8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 uint    __ovld __conv intel_sub_group_block_read( const __global uint* p );
 uint2   __ovld __conv intel_sub_group_block_read2( const __global uint* p );
@@ -15590,12 +15590,12 @@
 void    __ovld __conv intel_sub_group_block_write4(write_only image2d_t image, int2 coord, uint4 data);
 void    __ovld __conv intel_sub_group_block_write8(write_only image2d_t image, int2 coord, uint8 data);
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void    __ovld __conv intel_sub_group_block_write(read_write image2d_t image, int2 coord, uint data);
 void    __ovld __conv intel_sub_group_block_write2(read_write image2d_t image, int2 coord, uint2 data);
 void    __ovld __conv intel_sub_group_block_write4(read_write image2d_t image, int2 coord, uint4 data);
 void    __ovld __conv intel_sub_group_block_write8(read_write image2d_t image, int2 coord, uint8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void    __ovld __conv intel_sub_group_block_write( __global uint* p, uint data );
 void    __ovld __conv intel_sub_group_block_write2( __global uint* p, uint2 data );
@@ -15713,12 +15713,12 @@
 uint4      __ovld __conv intel_sub_group_block_read_ui4( read_only image2d_t image, int2 byte_coord );
 uint8      __ovld __conv intel_sub_group_block_read_ui8( read_only image2d_t image, int2 byte_coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint       __ovld __conv intel_sub_group_block_read_ui( read_write image2d_t image, int2 byte_coord );
 uint2      __ovld __conv intel_sub_group_block_read_ui2( read_write image2d_t image, int2 byte_coord );
 uint4      __ovld __conv intel_sub_group_block_read_ui4( read_write image2d_t image, int2 byte_coord );
 uint8      __ovld __conv intel_sub_group_block_read_ui8( read_write image2d_t image, int2 byte_coord );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 uint       __ovld __conv intel_sub_group_block_read_ui( const __global uint* p );
 uint2      __ovld __conv intel_sub_group_block_read_ui2( const __global uint* p );
@@ -15730,12 +15730,12 @@
 void       __ovld __conv intel_sub_group_block_write_ui4( read_only image2d_t image, int2 byte_coord, uint4 data );
 void       __ovld __conv intel_sub_group_block_write_ui8( read_only image2d_t image, int2 byte_coord, uint8 data );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void       __ovld __conv intel_sub_group_block_write_ui( read_write image2d_t image, int2 byte_coord, uint data );
 void       __ovld __conv intel_sub_group_block_write_ui2( read_write image2d_t image, int2 byte_coord, uint2 data );
 void       __ovld __conv intel_sub_group_block_write_ui4( read_write image2d_t image, int2 byte_coord, uint4 data );
 void       __ovld __conv intel_sub_group_block_write_ui8( read_write image2d_t image, int2 byte_coord, uint8 data );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void       __ovld __conv intel_sub_group_block_write_ui( __global uint* p, uint data );
 void       __ovld __conv intel_sub_group_block_write_ui2( __global uint* p, uint2 data );
@@ -15747,12 +15747,12 @@
 ushort4     __ovld __conv intel_sub_group_block_read_us4( read_only image2d_t image, int2 coord );
 ushort8     __ovld __conv intel_sub_group_block_read_us8( read_only image2d_t image, int2 coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 ushort      __ovld __conv intel_sub_group_block_read_us(read_write image2d_t image, int2 coord);
 ushort2     __ovld __conv intel_sub_group_block_read_us2(read_write image2d_t image, int2 coord);
 ushort4     __ovld __conv intel_sub_group_block_read_us4(read_write image2d_t image, int2 coord);
 ushort8     __ovld __conv intel_sub_group_block_read_us8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 ushort      __ovld __conv intel_sub_group_block_read_us(  const __global ushort* p );
 ushort2     __ovld __conv intel_sub_group_block_read_us2( const __global ushort* p );
@@ -15764,12 +15764,12 @@
 void        __ovld __conv intel_sub_group_block_write_us4(write_only image2d_t image, int2 coord, ushort4 data);
 void        __ovld __conv intel_sub_group_block_write_us8(write_only image2d_t image, int2 coord, ushort8 data);
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void        __ovld __conv intel_sub_group_block_write_us(read_write image2d_t image, int2 coord, ushort  data);
 void        __ovld __conv intel_sub_group_block_write_us2(read_write image2d_t image, int2 coord, ushort2 data);
 void        __ovld __conv intel_sub_group_block_write_us4(read_write image2d_t image, int2 coord, ushort4 data);
 void        __ovld __conv intel_sub_group_block_write_us8(read_write image2d_t image, int2 coord, ushort8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void        __ovld __conv intel_sub_group_block_write_us(  __global ushort* p, ushort  data );
 void        __ovld __conv intel_sub_group_block_write_us2( __global ushort* p, ushort2 data );
diff --git a/darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math.h b/darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math.h
rename to darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math_declares.h b/darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math_declares.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math_declares.h
rename to darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math_declares.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/cmath b/darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/cmath
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/cmath
rename to darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/cmath
diff --git a/darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/math.h b/darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/math.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/openmp_wrappers/math.h
rename to darwin-x86/lib64/clang/10.0.1/include/openmp_wrappers/math.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/pconfigintrin.h b/darwin-x86/lib64/clang/10.0.1/include/pconfigintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/pconfigintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/pconfigintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/pkuintrin.h b/darwin-x86/lib64/clang/10.0.1/include/pkuintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/pkuintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/pkuintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/pmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/pmmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/pmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/pmmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/popcntintrin.h b/darwin-x86/lib64/clang/10.0.1/include/popcntintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/popcntintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/popcntintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
index 617ce24..293276c 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
@@ -35,6 +35,8 @@
 #ifndef EMMINTRIN_H_
 #define EMMINTRIN_H_
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <altivec.h>
 
 /* We need definitions from the SSE header files.  */
@@ -2315,4 +2317,8 @@
   return (__m128d) __A;
 }
 
+#else
+#include_next <emmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* EMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
similarity index 84%
rename from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
rename to darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
index 3658919..24b14c8 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
@@ -10,6 +10,8 @@
 #ifndef _MM_MALLOC_H_INCLUDED
 #define _MM_MALLOC_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <stdlib.h>
 
 /* We can't depend on <stdlib.h> since the prototype of posix_memalign
@@ -25,12 +27,8 @@
 {
   /* PowerPC64 ELF V2 ABI requires quadword alignment.  */
   size_t vec_align = sizeof (__vector float);
-  /* Linux GLIBC malloc alignment is at least 2 X ptr size.  */
-  size_t malloc_align = (sizeof (void *) + sizeof (void *));
   void *ptr;
 
-  if (alignment == malloc_align && alignment == vec_align)
-    return malloc (size);
   if (alignment < vec_align)
     alignment = vec_align;
   if (posix_memalign (&ptr, alignment, size) == 0)
@@ -45,4 +43,8 @@
   free (ptr);
 }
 
+#else
+#include_next <mm_malloc.h>
+#endif
+
 #endif /* _MM_MALLOC_H_INCLUDED */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
index b949653..c55c447 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
@@ -35,6 +35,8 @@
 #ifndef _MMINTRIN_H_INCLUDED
 #define _MMINTRIN_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <altivec.h>
 /* The Intel API is flexible enough that we must allow aliasing with other
    vector types, and their scalar components.  */
@@ -1440,4 +1442,9 @@
   return (res.as_m64);
 #endif
 }
+
+#else
+#include_next <mmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* _MMINTRIN_H_INCLUDED */
diff --git a/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h
new file mode 100644
index 0000000..6d93383
--- /dev/null
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h
@@ -0,0 +1,150 @@
+/*===---- pmmintrin.h - Implementation of SSE3 intrinsics on PowerPC -------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerpc64le.
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.
+
+   In the specific case of X86 SSE3 intrinsics, the PowerPC VMX/VSX ISA
+   is a good match for most SIMD operations.  However the Horizontal
+   add/sub requires the data pairs be permuted into a separate
+   registers with vertical even/odd alignment for the operation.
+   And the addsub operation requires the sign of only the even numbered
+   elements be flipped (xored with -0.0).
+   For larger blocks of code using these intrinsic implementations,
+   the compiler be should be able to schedule instructions to avoid
+   additional latency.
+
+   In the specific case of the monitor and mwait instructions there are
+   no direct equivalent in the PowerISA at this time.  So those
+   intrinsics are not implemented.  */
+#error "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this warning."
+#endif
+
+#ifndef PMMINTRIN_H_
+#define PMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+/* We need definitions from the SSE2 and SSE header files*/
+#include <emmintrin.h>
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_addsub_ps (__m128 __X, __m128 __Y)
+{
+  const __v4sf even_n0 = {-0.0, 0.0, -0.0, 0.0};
+  __v4sf even_neg_Y = vec_xor(__Y, even_n0);
+  return (__m128) vec_add (__X, even_neg_Y);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_addsub_pd (__m128d __X, __m128d __Y)
+{
+  const __v2df even_n0 = {-0.0, 0.0};
+  __v2df even_neg_Y = vec_xor(__Y, even_n0);
+  return (__m128d) vec_add (__X, even_neg_Y);
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_ps (__m128 __X, __m128 __Y)
+{
+  __vector unsigned char xform2 = {
+      0x00, 0x01, 0x02, 0x03,
+      0x08, 0x09, 0x0A, 0x0B,
+      0x10, 0x11, 0x12, 0x13,
+      0x18, 0x19, 0x1A, 0x1B
+    };
+  __vector unsigned char xform1 = {
+      0x04, 0x05, 0x06, 0x07,
+      0x0C, 0x0D, 0x0E, 0x0F,
+      0x14, 0x15, 0x16, 0x17,
+      0x1C, 0x1D, 0x1E, 0x1F
+    };
+  return (__m128) vec_add (vec_perm ((__v4sf) __X, (__v4sf) __Y, xform2),
+			   vec_perm ((__v4sf) __X, (__v4sf) __Y, xform1));
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_ps (__m128 __X, __m128 __Y)
+{
+  __vector unsigned char xform2 = {
+      0x00, 0x01, 0x02, 0x03,
+      0x08, 0x09, 0x0A, 0x0B,
+      0x10, 0x11, 0x12, 0x13,
+      0x18, 0x19, 0x1A, 0x1B
+    };
+  __vector unsigned char xform1 = {
+      0x04, 0x05, 0x06, 0x07,
+      0x0C, 0x0D, 0x0E, 0x0F,
+      0x14, 0x15, 0x16, 0x17,
+      0x1C, 0x1D, 0x1E, 0x1F
+    };
+  return (__m128) vec_sub (vec_perm ((__v4sf) __X, (__v4sf) __Y, xform2),
+			   vec_perm ((__v4sf) __X, (__v4sf) __Y, xform1));
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pd (__m128d __X, __m128d __Y)
+{
+  return (__m128d) vec_add (vec_mergeh ((__v2df) __X, (__v2df)__Y),
+				  vec_mergel ((__v2df) __X, (__v2df)__Y));
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pd (__m128d __X, __m128d __Y)
+{
+  return (__m128d) vec_sub (vec_mergeh ((__v2df) __X, (__v2df)__Y),
+			    vec_mergel ((__v2df) __X, (__v2df)__Y));
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_movehdup_ps (__m128 __X)
+{
+  return (__m128)vec_mergeo ((__v4su)__X, (__v4su)__X);
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_moveldup_ps (__m128 __X)
+{
+  return (__m128)vec_mergee ((__v4su)__X, (__v4su)__X);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_loaddup_pd (double const *__P)
+{
+  return (__m128d) vec_splats (*__P);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_movedup_pd (__m128d __X)
+{
+  return _mm_shuffle_pd (__X, __X, _MM_SHUFFLE2 (0,0));
+}
+
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_lddqu_si128 (__m128i const *__P)
+{
+  return (__m128i) (vec_vsx_ld(0, (signed int const *)__P));
+}
+
+/* POWER8 / POWER9 have no equivalent for _mm_monitor nor _mm_wait.  */
+
+#else
+#include_next <pmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* PMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h
new file mode 100644
index 0000000..56ef6ba
--- /dev/null
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h
@@ -0,0 +1,85 @@
+/*===---- smmintrin.h - Implementation of SSE4 intrinsics on PowerPC -------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.
+
+   NOTE: This is NOT a complete implementation of the SSE4 intrinsics!  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerp64/powerpc64le.
+
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.  */
+#error                                                                         \
+    "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this error."
+#endif
+
+#ifndef SMMINTRIN_H_
+#define SMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+#include <altivec.h>
+#include <emmintrin.h>
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi8(__m128i __X, const int __N) {
+  return (unsigned char)((__v16qi)__X)[__N & 15];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi32(__m128i __X, const int __N) {
+  return ((__v4si)__X)[__N & 3];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi64(__m128i __X, const int __N) {
+  return ((__v2di)__X)[__N & 1];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_ps(__m128 __X, const int __N) {
+  return ((__v4si)__X)[__N & 3];
+}
+
+extern __inline __m128i
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_blend_epi16(__m128i __A, __m128i __B, const int __imm8) {
+  __v16qi __charmask = vec_splats((signed char)__imm8);
+  __charmask = vec_gb(__charmask);
+  __v8hu __shortmask = (__v8hu)vec_unpackh(__charmask);
+#ifdef __BIG_ENDIAN__
+  __shortmask = vec_reve(__shortmask);
+#endif
+  return (__m128i)vec_sel((__v8hu)__A, (__v8hu)__B, __shortmask);
+}
+
+extern __inline __m128i
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_blendv_epi8(__m128i __A, __m128i __B, __m128i __mask) {
+  const __v16qu __seven = vec_splats((unsigned char)0x07);
+  __v16qu __lmask = vec_sra((__v16qu)__mask, __seven);
+  return (__m128i)vec_sel((__v16qu)__A, (__v16qu)__B, __lmask);
+}
+
+#else
+#include_next <smmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* _SMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h
new file mode 100644
index 0000000..b5a935d
--- /dev/null
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h
@@ -0,0 +1,495 @@
+/*===---- tmmintrin.h - Implementation of SSSE3 intrinsics on PowerPC ------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerpc64le.
+
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.  */
+#endif
+
+#ifndef TMMINTRIN_H_
+#define TMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+#include <altivec.h>
+
+/* We need definitions from the SSE header files.  */
+#include <pmmintrin.h>
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi16 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v8hi) __A);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi32 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v4si) __A);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi8 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v16qi) __A);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi16 (__m64 __A)
+{
+  __v8hi __B = (__v8hi) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi32 (__m64 __A)
+{
+  __v4si __B = (__v4si) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi8 (__m64 __A)
+{
+  __v16qi __B = (__v16qi) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_alignr_epi8 (__m128i __A, __m128i __B, const unsigned int __count)
+{
+  if (__builtin_constant_p (__count) && __count < 16)
+    {
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_reve ((__v16qu) __A);
+      __B = (__m128i) vec_reve ((__v16qu) __B);
+#endif
+      __A = (__m128i) vec_sld ((__v16qu) __B, (__v16qu) __A, __count);
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_reve ((__v16qu) __A);
+#endif
+      return __A;
+    }
+
+  if (__count == 0)
+    return __B;
+
+  if (__count >= 16)
+    {
+      if (__count >= 32)
+	{
+	  const __v16qu zero = { 0 };
+	  return (__m128i) zero;
+	}
+      else
+	{
+	  const __v16qu __shift =
+	    vec_splats ((unsigned char) ((__count - 16) * 8));
+#ifdef __LITTLE_ENDIAN__
+	  return (__m128i) vec_sro ((__v16qu) __A, __shift);
+#else
+	  return (__m128i) vec_slo ((__v16qu) __A, __shift);
+#endif
+	}
+    }
+  else
+    {
+      const __v16qu __shiftA =
+	vec_splats ((unsigned char) ((16 - __count) * 8));
+      const __v16qu __shiftB = vec_splats ((unsigned char) (__count * 8));
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_slo ((__v16qu) __A, __shiftA);
+      __B = (__m128i) vec_sro ((__v16qu) __B, __shiftB);
+#else
+      __A = (__m128i) vec_sro ((__v16qu) __A, __shiftA);
+      __B = (__m128i) vec_slo ((__v16qu) __B, __shiftB);
+#endif
+      return (__m128i) vec_or ((__v16qu) __A, (__v16qu) __B);
+    }
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_alignr_pi8 (__m64 __A, __m64 __B, unsigned int __count)
+{
+  if (__count < 16)
+    {
+      __v2du __C = { __B, __A };
+#ifdef __LITTLE_ENDIAN__
+      const __v4su __shift = { __count << 3, 0, 0, 0 };
+      __C = (__v2du) vec_sro ((__v16qu) __C, (__v16qu) __shift);
+#else
+      const __v4su __shift = { 0, 0, 0, __count << 3 };
+      __C = (__v2du) vec_slo ((__v16qu) __C, (__v16qu) __shift);
+#endif
+      return (__m64) __C[0];
+    }
+  else
+    {
+      const __m64 __zero = { 0 };
+      return __zero;
+    }
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_add (__C, __D);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_epi32 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 };
+  __v4si __C = vec_perm ((__v4si) __A, (__v4si) __B, __P);
+  __v4si __D = vec_perm ((__v4si) __A, (__v4si) __B, __Q);
+  return (__m128i) vec_add (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pi16 (__m64 __A, __m64 __B)
+{
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_add (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pi32 (__m64 __A, __m64 __B)
+{
+  __v4si __C = (__v4si) (__v2du) { __A, __B };
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11,  0,  1,  2,  3,  8,  9, 10, 11 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15,  4,  5,  6,  7, 12, 13, 14, 15 };
+  __v4si __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_add (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadds_epi16 (__m128i __A, __m128i __B)
+{
+  __v4si __C = { 0 }, __D = { 0 };
+  __C = vec_sum4s ((__v8hi) __A, __C);
+  __D = vec_sum4s ((__v8hi) __B, __D);
+  __C = (__v4si) vec_packs (__C, __D);
+  return (__m128i) __C;
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadds_pi16 (__m64 __A, __m64 __B)
+{
+  const __v4si __zero = { 0 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v4si __D = vec_sum4s (__C, __zero);
+  __C = vec_packs (__D, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_sub (__C, __D);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_epi32 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 };
+  __v4si __C = vec_perm ((__v4si) __A, (__v4si) __B, __P);
+  __v4si __D = vec_perm ((__v4si) __A, (__v4si) __B, __Q);
+  return (__m128i) vec_sub (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pi16 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v8hi __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_sub (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pi32 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11,  0,  1,  2,  3,  8,  9, 10, 11 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15,  4,  5,  6,  7, 12, 13, 14, 15 };
+  __v4si __C = (__v4si) (__v2du) { __A, __B };
+  __v4si __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_sub (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsubs_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_subs (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsubs_pi16 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v8hi __D = vec_perm (__C, __C, __P);
+  __v8hi __E = vec_perm (__C, __C, __Q);
+  __C = vec_subs (__D, __E);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_shuffle_epi8 (__m128i __A, __m128i __B)
+{
+  const __v16qi __zero = { 0 };
+  __vector __bool char __select = vec_cmplt ((__v16qi) __B, __zero);
+  __v16qi __C = vec_perm ((__v16qi) __A, (__v16qi) __A, (__v16qu) __B);
+  return (__m128i) vec_sel (__C, __zero, __select);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_shuffle_pi8 (__m64 __A, __m64 __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __C = (__v16qi) (__v2du) { __A, __A };
+  __v16qi __D = (__v16qi) (__v2du) { __B, __B };
+  __vector __bool char __select = vec_cmplt ((__v16qi) __D, __zero);
+  __C = vec_perm ((__v16qi) __C, (__v16qi) __C, (__v16qu) __D);
+  __C = vec_sel (__C, __zero, __select);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi8 (__m128i __A, __m128i __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __selectneg = (__v16qi) vec_cmplt ((__v16qi) __B, __zero);
+  __v16qi __selectpos =
+    (__v16qi) vec_neg ((__v16qi) vec_cmpgt ((__v16qi) __B, __zero));
+  __v16qi __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v16qi) __A, (__v16qi) __conv);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi16 (__m128i __A, __m128i __B)
+{
+  const __v8hi __zero = { 0 };
+  __v8hi __selectneg = (__v8hi) vec_cmplt ((__v8hi) __B, __zero);
+  __v8hi __selectpos =
+    (__v8hi) vec_neg ((__v8hi) vec_cmpgt ((__v8hi) __B, __zero));
+  __v8hi __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v8hi) __A, (__v8hi) __conv);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi32 (__m128i __A, __m128i __B)
+{
+  const __v4si __zero = { 0 };
+  __v4si __selectneg = (__v4si) vec_cmplt ((__v4si) __B, __zero);
+  __v4si __selectpos =
+    (__v4si) vec_neg ((__v4si) vec_cmpgt ((__v4si) __B, __zero));
+  __v4si __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v4si) __A, (__v4si) __conv);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi8 (__m64 __A, __m64 __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __C = (__v16qi) (__v2du) { __A, __A };
+  __v16qi __D = (__v16qi) (__v2du) { __B, __B };
+  __C = (__v16qi) _mm_sign_epi8 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi16 (__m64 __A, __m64 __B)
+{
+  const __v8hi __zero = { 0 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __A };
+  __v8hi __D = (__v8hi) (__v2du) { __B, __B };
+  __C = (__v8hi) _mm_sign_epi16 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi32 (__m64 __A, __m64 __B)
+{
+  const __v4si __zero = { 0 };
+  __v4si __C = (__v4si) (__v2du) { __A, __A };
+  __v4si __D = (__v4si) (__v2du) { __B, __B };
+  __C = (__v4si) _mm_sign_epi32 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maddubs_epi16 (__m128i __A, __m128i __B)
+{
+  __v8hi __unsigned = vec_splats ((signed short) 0x00ff);
+  __v8hi __C = vec_and (vec_unpackh ((__v16qi) __A), __unsigned);
+  __v8hi __D = vec_and (vec_unpackl ((__v16qi) __A), __unsigned);
+  __v8hi __E = vec_unpackh ((__v16qi) __B);
+  __v8hi __F = vec_unpackl ((__v16qi) __B);
+  __C = vec_mul (__C, __E);
+  __D = vec_mul (__D, __F);
+  const __v16qu __odds  =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __evens =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __E = vec_perm (__C, __D, __odds);
+  __F = vec_perm (__C, __D, __evens);
+  return (__m128i) vec_adds (__E, __F);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maddubs_pi16 (__m64 __A, __m64 __B)
+{
+  __v8hi __C = (__v8hi) (__v2du) { __A, __A };
+  __C = vec_unpackl ((__v16qi) __C);
+  const __v8hi __unsigned = vec_splats ((signed short) 0x00ff);
+  __C = vec_and (__C, __unsigned);
+  __v8hi __D = (__v8hi) (__v2du) { __B, __B };
+  __D = vec_unpackl ((__v16qi) __D);
+  __D = vec_mul (__C, __D);
+  const __v16qu __odds  =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __evens =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __C = vec_perm (__D, __D, __odds);
+  __D = vec_perm (__D, __D, __evens);
+  __C = vec_adds (__C, __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mulhrs_epi16 (__m128i __A, __m128i __B)
+{
+  __v4si __C = vec_unpackh ((__v8hi) __A);
+  __v4si __D = vec_unpackh ((__v8hi) __B);
+  __C = vec_mul (__C, __D);
+  __D = vec_unpackl ((__v8hi) __A);
+  __v4si __E = vec_unpackl ((__v8hi) __B);
+  __D = vec_mul (__D, __E);
+  const __v4su __shift = vec_splats ((unsigned int) 14);
+  __C = vec_sr (__C, __shift);
+  __D = vec_sr (__D, __shift);
+  const __v4si __ones = vec_splats ((signed int) 1);
+  __C = vec_add (__C, __ones);
+  __C = vec_sr (__C, (__v4su) __ones);
+  __D = vec_add (__D, __ones);
+  __D = vec_sr (__D, (__v4su) __ones);
+  return (__m128i) vec_pack (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mulhrs_pi16 (__m64 __A, __m64 __B)
+{
+  __v4si __C = (__v4si) (__v2du) { __A, __A };
+  __C = vec_unpackh ((__v8hi) __C);
+  __v4si __D = (__v4si) (__v2du) { __B, __B };
+  __D = vec_unpackh ((__v8hi) __D);
+  __C = vec_mul (__C, __D);
+  const __v4su __shift = vec_splats ((unsigned int) 14);
+  __C = vec_sr (__C, __shift);
+  const __v4si __ones = vec_splats ((signed int) 1);
+  __C = vec_add (__C, __ones);
+  __C = vec_sr (__C, (__v4su) __ones);
+  __v8hi __E = vec_pack (__C, __D);
+  return (__m64) ((__v2du) (__E))[0];
+}
+
+#else
+#include_next <tmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* TMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
index 1b322b6..0f429fa 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
@@ -34,6 +34,8 @@
 #ifndef _XMMINTRIN_H_INCLUDED
 #define _XMMINTRIN_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 /* Define four value permute mask */
 #define _MM_SHUFFLE(w,x,y,z) (((w) << 6) | ((x) << 4) | ((y) << 2) | (z))
 
@@ -1835,4 +1837,8 @@
 /* For backward source compatibility.  */
 //# include <emmintrin.h>
 
+#else
+#include_next <xmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* _XMMINTRIN_H_INCLUDED */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/prfchwintrin.h b/darwin-x86/lib64/clang/10.0.1/include/prfchwintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/prfchwintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/prfchwintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ptwriteintrin.h b/darwin-x86/lib64/clang/10.0.1/include/ptwriteintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/ptwriteintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/ptwriteintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/rdseedintrin.h b/darwin-x86/lib64/clang/10.0.1/include/rdseedintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/rdseedintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/rdseedintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/rtmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/rtmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/rtmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/rtmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/s390intrin.h b/darwin-x86/lib64/clang/10.0.1/include/s390intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/s390intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/s390intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/allocator_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/allocator_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/allocator_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/allocator_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/asan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/asan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/asan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/asan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/common_interface_defs.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/common_interface_defs.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/common_interface_defs.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/common_interface_defs.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/coverage_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/coverage_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/coverage_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/coverage_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/dfsan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/dfsan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/dfsan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/dfsan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/hwasan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/hwasan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/hwasan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/hwasan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/linux_syscall_hooks.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/linux_syscall_hooks.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/linux_syscall_hooks.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/linux_syscall_hooks.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/lsan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/lsan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/lsan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/lsan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/msan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/msan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/msan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/msan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/netbsd_syscall_hooks.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/netbsd_syscall_hooks.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/netbsd_syscall_hooks.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/netbsd_syscall_hooks.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/scudo_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/scudo_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/scudo_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/scudo_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface_atomic.h b/darwin-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface_atomic.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface_atomic.h
rename to darwin-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface_atomic.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/sgxintrin.h b/darwin-x86/lib64/clang/10.0.1/include/sgxintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/sgxintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/sgxintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/shaintrin.h b/darwin-x86/lib64/clang/10.0.1/include/shaintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/shaintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/shaintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/smmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/smmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/smmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/smmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdalign.h b/darwin-x86/lib64/clang/10.0.1/include/stdalign.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdalign.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdalign.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdarg.h b/darwin-x86/lib64/clang/10.0.1/include/stdarg.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdarg.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdarg.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdatomic.h b/darwin-x86/lib64/clang/10.0.1/include/stdatomic.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdatomic.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdatomic.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdbool.h b/darwin-x86/lib64/clang/10.0.1/include/stdbool.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdbool.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdbool.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stddef.h b/darwin-x86/lib64/clang/10.0.1/include/stddef.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stddef.h
rename to darwin-x86/lib64/clang/10.0.1/include/stddef.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdint.h b/darwin-x86/lib64/clang/10.0.1/include/stdint.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdint.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdint.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/stdnoreturn.h b/darwin-x86/lib64/clang/10.0.1/include/stdnoreturn.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/stdnoreturn.h
rename to darwin-x86/lib64/clang/10.0.1/include/stdnoreturn.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/tbmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/tbmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/tbmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/tbmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/tgmath.h b/darwin-x86/lib64/clang/10.0.1/include/tgmath.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/tgmath.h
rename to darwin-x86/lib64/clang/10.0.1/include/tgmath.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/tmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/tmmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/tmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/tmmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/unwind.h b/darwin-x86/lib64/clang/10.0.1/include/unwind.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/unwind.h
rename to darwin-x86/lib64/clang/10.0.1/include/unwind.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/vadefs.h b/darwin-x86/lib64/clang/10.0.1/include/vadefs.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/vadefs.h
rename to darwin-x86/lib64/clang/10.0.1/include/vadefs.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/vaesintrin.h b/darwin-x86/lib64/clang/10.0.1/include/vaesintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/vaesintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/vaesintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/varargs.h b/darwin-x86/lib64/clang/10.0.1/include/varargs.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/varargs.h
rename to darwin-x86/lib64/clang/10.0.1/include/varargs.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/vecintrin.h b/darwin-x86/lib64/clang/10.0.1/include/vecintrin.h
similarity index 95%
rename from darwin-x86/lib64/clang/9.0.7/include/vecintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/vecintrin.h
index 115d893..c71b76a 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/vecintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/vecintrin.h
@@ -517,6 +517,141 @@
 }
 #endif
 
+/*-- vec_revb ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_revb(vector signed short __vec) {
+  return (vector signed short)
+         __builtin_s390_vlbrh((vector unsigned short)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_revb(vector unsigned short __vec) {
+  return __builtin_s390_vlbrh(__vec);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_revb(vector signed int __vec) {
+  return (vector signed int)
+         __builtin_s390_vlbrf((vector unsigned int)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_revb(vector unsigned int __vec) {
+  return __builtin_s390_vlbrf(__vec);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_revb(vector signed long long __vec) {
+  return (vector signed long long)
+         __builtin_s390_vlbrg((vector unsigned long long)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_revb(vector unsigned long long __vec) {
+  return __builtin_s390_vlbrg(__vec);
+}
+
+#if __ARCH__ >= 12
+static inline __ATTRS_o_ai vector float
+vec_revb(vector float __vec) {
+  return (vector float)
+         __builtin_s390_vlbrf((vector unsigned int)__vec);
+}
+#endif
+
+static inline __ATTRS_o_ai vector double
+vec_revb(vector double __vec) {
+  return (vector double)
+         __builtin_s390_vlbrg((vector unsigned long long)__vec);
+}
+
+/*-- vec_reve ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_reve(vector signed char __vec) {
+  return (vector signed char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                                __vec[11], __vec[10], __vec[9], __vec[8],
+                                __vec[7], __vec[6], __vec[5], __vec[4],
+                                __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_reve(vector unsigned char __vec) {
+  return (vector unsigned char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                                  __vec[11], __vec[10], __vec[9], __vec[8],
+                                  __vec[7], __vec[6], __vec[5], __vec[4],
+                                  __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_reve(vector bool char __vec) {
+  return (vector bool char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                              __vec[11], __vec[10], __vec[9], __vec[8],
+                              __vec[7], __vec[6], __vec[5], __vec[4],
+                              __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_reve(vector signed short __vec) {
+  return (vector signed short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                                 __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_reve(vector unsigned short __vec) {
+  return (vector unsigned short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                                   __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_reve(vector bool short __vec) {
+  return (vector bool short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                               __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_reve(vector signed int __vec) {
+  return (vector signed int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_reve(vector unsigned int __vec) {
+  return (vector unsigned int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_reve(vector bool int __vec) {
+  return (vector bool int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_reve(vector signed long long __vec) {
+  return (vector signed long long) { __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_reve(vector unsigned long long __vec) {
+  return (vector unsigned long long) { __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_reve(vector bool long long __vec) {
+  return (vector bool long long) { __vec[1], __vec[0] };
+}
+
+#if __ARCH__ >= 12
+static inline __ATTRS_o_ai vector float
+vec_reve(vector float __vec) {
+  return (vector float) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+#endif
+
+static inline __ATTRS_o_ai vector double
+vec_reve(vector double __vec) {
+  return (vector double) { __vec[1], __vec[0] };
+}
+
 /*-- vec_sel ----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -6835,6 +6970,56 @@
   __builtin_s390_vsldb((vector unsigned char)(X), \
                        (vector unsigned char)(Y), (Z) * 4))
 
+/*-- vec_sldb ---------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+extern __ATTRS_o vector signed char
+vec_sldb(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned char
+vec_sldb(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed short
+vec_sldb(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned short
+vec_sldb(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed int
+vec_sldb(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned int
+vec_sldb(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed long long
+vec_sldb(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned long long
+vec_sldb(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector float
+vec_sldb(vector float __a, vector float __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector double
+vec_sldb(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 7);
+
+#define vec_sldb(X, Y, Z) ((__typeof__((vec_sldb)((X), (Y), (Z)))) \
+  __builtin_s390_vsld((vector unsigned char)(X), \
+                      (vector unsigned char)(Y), (Z)))
+
+#endif
+
 /*-- vec_sral ---------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -7565,6 +7750,56 @@
     (vector unsigned char)__a, (vector unsigned char)__b);
 }
 
+/*-- vec_srdb ---------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+extern __ATTRS_o vector signed char
+vec_srdb(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned char
+vec_srdb(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed short
+vec_srdb(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned short
+vec_srdb(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed int
+vec_srdb(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned int
+vec_srdb(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed long long
+vec_srdb(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned long long
+vec_srdb(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector float
+vec_srdb(vector float __a, vector float __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector double
+vec_srdb(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 7);
+
+#define vec_srdb(X, Y, Z) ((__typeof__((vec_srdb)((X), (Y), (Z)))) \
+  __builtin_s390_vsrd((vector unsigned char)(X), \
+                      (vector unsigned char)(Y), (Z)))
+
+#endif
+
 /*-- vec_abs ----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -8711,6 +8946,22 @@
   return __builtin_convertvector(__a, vector double);
 }
 
+/*-- vec_float --------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector float
+vec_float(vector signed int __a) {
+  return __builtin_convertvector(__a, vector float);
+}
+
+static inline __ATTRS_o_ai vector float
+vec_float(vector unsigned int __a) {
+  return __builtin_convertvector(__a, vector float);
+}
+
+#endif
+
 /*-- vec_signed -------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed long long
@@ -8718,6 +8969,13 @@
   return __builtin_convertvector(__a, vector signed long long);
 }
 
+#if __ARCH__ >= 13
+static inline __ATTRS_o_ai vector signed int
+vec_signed(vector float __a) {
+  return __builtin_convertvector(__a, vector signed int);
+}
+#endif
+
 /*-- vec_unsigned -----------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector unsigned long long
@@ -8725,6 +8983,13 @@
   return __builtin_convertvector(__a, vector unsigned long long);
 }
 
+#if __ARCH__ >= 13
+static inline __ATTRS_o_ai vector unsigned int
+vec_unsigned(vector float __a) {
+  return __builtin_convertvector(__a, vector unsigned int);
+}
+#endif
+
 /*-- vec_roundp -------------------------------------------------------------*/
 
 #if __ARCH__ >= 12
@@ -10442,6 +10707,147 @@
   return __builtin_s390_vfaezfs(__a, __b, 8, __cc);
 }
 
+/*-- vec_search_string_cc ---------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed char __a, vector signed char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb((vector unsigned char)__a,
+                               (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool char __a, vector bool char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb((vector unsigned char)__a,
+                               (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned char __a, vector unsigned char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed short __a, vector signed short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh((vector unsigned short)__a,
+                               (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool short __a, vector bool short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh((vector unsigned short)__a,
+                               (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned short __a, vector unsigned short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed int __a, vector signed int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf((vector unsigned int)__a,
+                               (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool int __a, vector bool int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf((vector unsigned int)__a,
+                               (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned int __a, vector unsigned int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf(__a, __b, __c, __cc);
+}
+
+#endif
+
+/*-- vec_search_string_until_zero_cc ----------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed char __a,
+                                vector signed char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb((vector unsigned char)__a,
+                                (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool char __a,
+                                vector bool char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb((vector unsigned char)__a,
+                                (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned char __a,
+                                vector unsigned char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed short __a,
+                                vector signed short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh((vector unsigned short)__a,
+                                (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool short __a,
+                                vector bool short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh((vector unsigned short)__a,
+                                (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned short __a,
+                                vector unsigned short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed int __a,
+                                vector signed int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf((vector unsigned int)__a,
+                                (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool int __a,
+                                vector bool int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf((vector unsigned int)__a,
+                                (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned int __a,
+                                vector unsigned int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf(__a, __b, __c, __cc);
+}
+
+#endif
+
 #undef __constant_pow2_range
 #undef __constant_range
 #undef __constant
diff --git a/darwin-x86/lib64/clang/9.0.7/include/vpclmulqdqintrin.h b/darwin-x86/lib64/clang/10.0.1/include/vpclmulqdqintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/vpclmulqdqintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/vpclmulqdqintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/waitpkgintrin.h b/darwin-x86/lib64/clang/10.0.1/include/waitpkgintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/waitpkgintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/waitpkgintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/wbnoinvdintrin.h b/darwin-x86/lib64/clang/10.0.1/include/wbnoinvdintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/wbnoinvdintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/wbnoinvdintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/wmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/wmmintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/wmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/wmmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/x86intrin.h b/darwin-x86/lib64/clang/10.0.1/include/x86intrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/x86intrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/x86intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xmmintrin.h
similarity index 99%
rename from darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xmmintrin.h
index 6f5517e..75ff376 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h
+++ b/darwin-x86/lib64/clang/10.0.1/include/xmmintrin.h
@@ -1919,7 +1919,11 @@
 static __inline__ void __DEFAULT_FN_ATTRS
 _mm_storeh_pi(__m64 *__p, __m128 __a)
 {
-  __builtin_ia32_storehps((__v2si *)__p, (__v4sf)__a);
+  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_storeh_pi_struct {
+    __mm_storeh_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 2, 3);
 }
 
 /// Stores the lower 64 bits of a 128-bit vector of [4 x float] to a
@@ -1936,7 +1940,11 @@
 static __inline__ void __DEFAULT_FN_ATTRS
 _mm_storel_pi(__m64 *__p, __m128 __a)
 {
-  __builtin_ia32_storelps((__v2si *)__p, (__v4sf)__a);
+  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_storeh_pi_struct {
+    __mm_storeh_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 0, 1);
 }
 
 /// Stores the lower 32 bits of a 128-bit vector of [4 x float] to a
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xopintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xopintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xopintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xopintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xsavecintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xsavecintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xsavecintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xsavecintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xsaveintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xsaveintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xsaveintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xsaveintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xsaveoptintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xsaveoptintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xsaveoptintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xsaveoptintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xsavesintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xsavesintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xsavesintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xsavesintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xtestintrin.h b/darwin-x86/lib64/clang/10.0.1/include/xtestintrin.h
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/include/xtestintrin.h
rename to darwin-x86/lib64/clang/10.0.1/include/xtestintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/share/asan_blacklist.txt b/darwin-x86/lib64/clang/10.0.1/share/asan_blacklist.txt
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/share/asan_blacklist.txt
rename to darwin-x86/lib64/clang/10.0.1/share/asan_blacklist.txt
diff --git a/darwin-x86/lib64/clang/9.0.7/share/cfi_blacklist.txt b/darwin-x86/lib64/clang/10.0.1/share/cfi_blacklist.txt
similarity index 100%
rename from darwin-x86/lib64/clang/9.0.7/share/cfi_blacklist.txt
rename to darwin-x86/lib64/clang/10.0.1/share/cfi_blacklist.txt
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm_fp16.h b/darwin-x86/lib64/clang/9.0.7/include/arm_fp16.h
deleted file mode 100644
index de54465..0000000
--- a/darwin-x86/lib64/clang/9.0.7/include/arm_fp16.h
+++ /dev/null
@@ -1,1499 +0,0 @@
-/*===---- arm_fp16.h - ARM FP16 intrinsics ---------------------------------===
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
- * THE SOFTWARE.
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __ARM_FP16_H
-#define __ARM_FP16_H
-
-#include <stdint.h>
-
-typedef __fp16 float16_t;
-#define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
-
-#if defined(__ARM_FEATURE_FP16_SCALAR_ARITHMETIC) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vabdh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vabdh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vabsh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vabsh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vaddh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vaddh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcageh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcageh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcagth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcagth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcaleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcaleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcalth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcalth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vceqh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vceqh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vceqzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vceqzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgeh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcgeh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcgezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcgth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgtzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcgtzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vclezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vclezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vclth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vclth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcltzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcltzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u32(uint32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u32(uint32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u64(uint64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u64(uint64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u16(uint16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u16(uint16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s32(int32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s32(int32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s64(int64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s64(int64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s16(int16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s16(int16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdivh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vdivh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#else
-#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#else
-#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmaxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vminh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vminnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmulh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmulxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vnegh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vnegh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpeh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrecpeh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrecpxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndah_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndah_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndih_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndih_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndmh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndmh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndnh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndnh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndph_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndph_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsqrteh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrsqrteh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsqrth_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
-  __ret; \
-})
-#else
-#define vsqrth_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsubh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vsubh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#endif
-
-#undef __ai
-
-#endif /* __ARM_FP16_H */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ompt.h b/darwin-x86/lib64/clang/9.0.7/include/ompt.h
deleted file mode 100644
index d04a2b7..0000000
--- a/darwin-x86/lib64/clang/9.0.7/include/ompt.h
+++ /dev/null
@@ -1,1082 +0,0 @@
-/*
- * include/50/omp-tools.h.var
- */
-
-//===----------------------------------------------------------------------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef __OMPT__
-#define __OMPT__
-
-/*****************************************************************************
- * system include files
- *****************************************************************************/
-
-#include <stdint.h>
-#include <stddef.h>
-
-/*****************************************************************************
- * iteration macros
- *****************************************************************************/
-
-#define FOREACH_OMPT_INQUIRY_FN(macro)      \
-    macro (ompt_enumerate_states)           \
-    macro (ompt_enumerate_mutex_impls)      \
-                                            \
-    macro (ompt_set_callback)               \
-    macro (ompt_get_callback)               \
-                                            \
-    macro (ompt_get_state)                  \
-                                            \
-    macro (ompt_get_parallel_info)          \
-    macro (ompt_get_task_info)              \
-    macro (ompt_get_task_memory)            \
-    macro (ompt_get_thread_data)            \
-    macro (ompt_get_unique_id)              \
-    macro (ompt_finalize_tool)              \
-                                            \
-    macro(ompt_get_num_procs)               \
-    macro(ompt_get_num_places)              \
-    macro(ompt_get_place_proc_ids)          \
-    macro(ompt_get_place_num)               \
-    macro(ompt_get_partition_place_nums)    \
-    macro(ompt_get_proc_id)                 \
-                                            \
-    macro(ompt_get_target_info)             \
-    macro(ompt_get_num_devices)
-
-#define FOREACH_OMPT_STATE(macro)                                                                \
-                                                                                                \
-    /* first available state */                                                                 \
-    macro (ompt_state_undefined, 0x102)      /* undefined thread state */                        \
-                                                                                                \
-    /* work states (0..15) */                                                                   \
-    macro (ompt_state_work_serial, 0x000)    /* working outside parallel */                      \
-    macro (ompt_state_work_parallel, 0x001)  /* working within parallel */                       \
-    macro (ompt_state_work_reduction, 0x002) /* performing a reduction */                        \
-                                                                                                \
-    /* barrier wait states (16..31) */                                                          \
-    macro (ompt_state_wait_barrier, 0x010)   /* waiting at a barrier */                          \
-    macro (ompt_state_wait_barrier_implicit_parallel, 0x011)                                     \
-                                            /* implicit barrier at the end of parallel region */\
-    macro (ompt_state_wait_barrier_implicit_workshare, 0x012)                                    \
-                                            /* implicit barrier at the end of worksharing */    \
-    macro (ompt_state_wait_barrier_implicit, 0x013)  /* implicit barrier */                      \
-    macro (ompt_state_wait_barrier_explicit, 0x014)  /* explicit barrier */                      \
-                                                                                                \
-    /* task wait states (32..63) */                                                             \
-    macro (ompt_state_wait_taskwait, 0x020)  /* waiting at a taskwait */                         \
-    macro (ompt_state_wait_taskgroup, 0x021) /* waiting at a taskgroup */                        \
-                                                                                                \
-    /* mutex wait states (64..127) */                                                           \
-    macro (ompt_state_wait_mutex, 0x040)                                                         \
-    macro (ompt_state_wait_lock, 0x041)      /* waiting for lock */                              \
-    macro (ompt_state_wait_critical, 0x042)  /* waiting for critical */                          \
-    macro (ompt_state_wait_atomic, 0x043)    /* waiting for atomic */                            \
-    macro (ompt_state_wait_ordered, 0x044)   /* waiting for ordered */                           \
-                                                                                                \
-    /* target wait states (128..255) */                                                         \
-    macro (ompt_state_wait_target, 0x080)        /* waiting for target region */                 \
-    macro (ompt_state_wait_target_map, 0x081)    /* waiting for target data mapping operation */ \
-    macro (ompt_state_wait_target_update, 0x082) /* waiting for target update operation */       \
-                                                                                                \
-    /* misc (256..511) */                                                                       \
-    macro (ompt_state_idle, 0x100)           /* waiting for work */                              \
-    macro (ompt_state_overhead, 0x101)       /* overhead excluding wait states */                \
-                                                                                                \
-    /* implementation-specific states (512..) */
-
-
-#define FOREACH_KMP_MUTEX_IMPL(macro)                                                \
-    macro (kmp_mutex_impl_none, 0)         /* unknown implementation */              \
-    macro (kmp_mutex_impl_spin, 1)         /* based on spin */                       \
-    macro (kmp_mutex_impl_queuing, 2)      /* based on some fair policy */           \
-    macro (kmp_mutex_impl_speculative, 3)  /* based on HW-supported speculation */
-
-#define FOREACH_OMPT_EVENT(macro)                                                                                        \
-                                                                                                                         \
-    /*--- Mandatory Events ---*/                                                                                         \
-    macro (ompt_callback_thread_begin,      ompt_callback_thread_begin_t,       1) /* thread begin                    */ \
-    macro (ompt_callback_thread_end,        ompt_callback_thread_end_t,         2) /* thread end                      */ \
-                                                                                                                         \
-    macro (ompt_callback_parallel_begin,    ompt_callback_parallel_begin_t,     3) /* parallel begin                  */ \
-    macro (ompt_callback_parallel_end,      ompt_callback_parallel_end_t,       4) /* parallel end                    */ \
-                                                                                                                         \
-    macro (ompt_callback_task_create,       ompt_callback_task_create_t,        5) /* task begin                      */ \
-    macro (ompt_callback_task_schedule,     ompt_callback_task_schedule_t,      6) /* task schedule                   */ \
-    macro (ompt_callback_implicit_task,     ompt_callback_implicit_task_t,      7) /* implicit task                   */ \
-                                                                                                                         \
-    macro (ompt_callback_target,            ompt_callback_target_t,             8) /* target                          */ \
-    macro (ompt_callback_target_data_op,    ompt_callback_target_data_op_t,     9) /* target data op                  */ \
-    macro (ompt_callback_target_submit,     ompt_callback_target_submit_t,     10) /* target  submit                  */ \
-                                                                                                                         \
-    macro (ompt_callback_control_tool,      ompt_callback_control_tool_t,      11) /* control tool                    */ \
-                                                                                                                         \
-    macro (ompt_callback_device_initialize, ompt_callback_device_initialize_t, 12) /* device initialize               */ \
-    macro (ompt_callback_device_finalize,   ompt_callback_device_finalize_t,   13) /* device finalize                 */ \
-                                                                                                                         \
-    macro (ompt_callback_device_load,       ompt_callback_device_load_t,       14) /* device load                     */ \
-    macro (ompt_callback_device_unload,     ompt_callback_device_unload_t,     15) /* device unload                   */ \
-                                                                                                                         \
-    /* Optional Events */                                                                                                \
-    macro (ompt_callback_sync_region_wait,  ompt_callback_sync_region_t,       16) /* sync region wait begin or end   */ \
-                                                                                                                         \
-    macro (ompt_callback_mutex_released,    ompt_callback_mutex_t,             17) /* mutex released                  */ \
-                                                                                                                         \
-    macro (ompt_callback_dependences,       ompt_callback_dependences_t,       18) /* report task dependences         */ \
-    macro (ompt_callback_task_dependence,   ompt_callback_task_dependence_t,   19) /* report task dependence          */ \
-                                                                                                                         \
-    macro (ompt_callback_work,              ompt_callback_work_t,              20) /* task at work begin or end       */ \
-                                                                                                                         \
-    macro (ompt_callback_master,            ompt_callback_master_t,            21) /* task at master begin or end     */ \
-                                                                                                                         \
-    macro (ompt_callback_target_map,        ompt_callback_target_map_t,        22) /* target map                      */ \
-                                                                                                                         \
-    macro (ompt_callback_sync_region,       ompt_callback_sync_region_t,       23) /* sync region begin or end        */ \
-                                                                                                                         \
-    macro (ompt_callback_lock_init,         ompt_callback_mutex_acquire_t,     24) /* lock init                       */ \
-    macro (ompt_callback_lock_destroy,      ompt_callback_mutex_t,             25) /* lock destroy                    */ \
-                                                                                                                         \
-    macro (ompt_callback_mutex_acquire,     ompt_callback_mutex_acquire_t,     26) /* mutex acquire                   */ \
-    macro (ompt_callback_mutex_acquired,    ompt_callback_mutex_t,             27) /* mutex acquired                  */ \
-                                                                                                                         \
-    macro (ompt_callback_nest_lock,         ompt_callback_nest_lock_t,         28) /* nest lock                       */ \
-                                                                                                                         \
-    macro (ompt_callback_flush,             ompt_callback_flush_t,             29) /* after executing flush           */ \
-                                                                                                                         \
-    macro (ompt_callback_cancel,            ompt_callback_cancel_t,            30) /* cancel innermost binding region */ \
-                                                                                                                         \
-    macro (ompt_callback_reduction,         ompt_callback_sync_region_t,       31) /* reduction                       */ \
-                                                                                                                         \
-    macro (ompt_callback_dispatch,          ompt_callback_dispatch_t,          32) /* dispatch of work                */
-
-/*****************************************************************************
- * implementation specific types
- *****************************************************************************/
-
-typedef enum kmp_mutex_impl_t {
-#define kmp_mutex_impl_macro(impl, code) impl = code,
-    FOREACH_KMP_MUTEX_IMPL(kmp_mutex_impl_macro)
-#undef kmp_mutex_impl_macro
-} kmp_mutex_impl_t;
-
-/*****************************************************************************
- * definitions generated from spec
- *****************************************************************************/
-
-typedef enum ompt_callbacks_t {
-  ompt_callback_thread_begin             = 1,
-  ompt_callback_thread_end               = 2,
-  ompt_callback_parallel_begin           = 3,
-  ompt_callback_parallel_end             = 4,
-  ompt_callback_task_create              = 5,
-  ompt_callback_task_schedule            = 6,
-  ompt_callback_implicit_task            = 7,
-  ompt_callback_target                   = 8,
-  ompt_callback_target_data_op           = 9,
-  ompt_callback_target_submit            = 10,
-  ompt_callback_control_tool             = 11,
-  ompt_callback_device_initialize        = 12,
-  ompt_callback_device_finalize          = 13,
-  ompt_callback_device_load              = 14,
-  ompt_callback_device_unload            = 15,
-  ompt_callback_sync_region_wait         = 16,
-  ompt_callback_mutex_released           = 17,
-  ompt_callback_dependences              = 18,
-  ompt_callback_task_dependence          = 19,
-  ompt_callback_work                     = 20,
-  ompt_callback_master                   = 21,
-  ompt_callback_target_map               = 22,
-  ompt_callback_sync_region              = 23,
-  ompt_callback_lock_init                = 24,
-  ompt_callback_lock_destroy             = 25,
-  ompt_callback_mutex_acquire            = 26,
-  ompt_callback_mutex_acquired           = 27,
-  ompt_callback_nest_lock                = 28,
-  ompt_callback_flush                    = 29,
-  ompt_callback_cancel                   = 30,
-  ompt_callback_reduction                = 31,
-  ompt_callback_dispatch                 = 32
-} ompt_callbacks_t;
-
-typedef enum ompt_record_t {
-  ompt_record_ompt               = 1,
-  ompt_record_native             = 2,
-  ompt_record_invalid            = 3
-} ompt_record_t;
-
-typedef enum ompt_record_native_t {
-  ompt_record_native_info  = 1,
-  ompt_record_native_event = 2
-} ompt_record_native_t;
-
-typedef enum ompt_set_result_t {
-  ompt_set_error            = 0,
-  ompt_set_never            = 1,
-  ompt_set_impossible       = 2,
-  ompt_set_sometimes        = 3,
-  ompt_set_sometimes_paired = 4,
-  ompt_set_always           = 5
-} ompt_set_result_t;
-
-typedef uint64_t ompt_id_t;
-
-typedef uint64_t ompt_device_time_t;
-
-typedef uint64_t ompt_buffer_cursor_t;
-
-typedef enum ompt_thread_t {
-  ompt_thread_initial                 = 1,
-  ompt_thread_worker                  = 2,
-  ompt_thread_other                   = 3,
-  ompt_thread_unknown                 = 4
-} ompt_thread_t;
-
-typedef enum ompt_scope_endpoint_t {
-  ompt_scope_begin                    = 1,
-  ompt_scope_end                      = 2
-} ompt_scope_endpoint_t;
-
-typedef enum ompt_dispatch_t {
-  ompt_dispatch_iteration             = 1,
-  ompt_dispatch_section               = 2
-} ompt_dispatch_t;
-
-typedef enum ompt_sync_region_t {
-  ompt_sync_region_barrier                = 1,
-  ompt_sync_region_barrier_implicit       = 2,
-  ompt_sync_region_barrier_explicit       = 3,
-  ompt_sync_region_barrier_implementation = 4,
-  ompt_sync_region_taskwait               = 5,
-  ompt_sync_region_taskgroup              = 6,
-  ompt_sync_region_reduction              = 7
-} ompt_sync_region_t;
-
-typedef enum ompt_target_data_op_t {
-  ompt_target_data_alloc                = 1,
-  ompt_target_data_transfer_to_device   = 2,
-  ompt_target_data_transfer_from_device = 3,
-  ompt_target_data_delete               = 4,
-  ompt_target_data_associate            = 5,
-  ompt_target_data_disassociate         = 6
-} ompt_target_data_op_t;
-
-typedef enum ompt_work_t {
-  ompt_work_loop               = 1,
-  ompt_work_sections           = 2,
-  ompt_work_single_executor    = 3,
-  ompt_work_single_other       = 4,
-  ompt_work_workshare          = 5,
-  ompt_work_distribute         = 6,
-  ompt_work_taskloop           = 7
-} ompt_work_t;
-
-typedef enum ompt_mutex_t {
-  ompt_mutex_lock                     = 1,
-  ompt_mutex_test_lock                = 2,
-  ompt_mutex_nest_lock                = 3,
-  ompt_mutex_test_nest_lock           = 4,
-  ompt_mutex_critical                 = 5,
-  ompt_mutex_atomic                   = 6,
-  ompt_mutex_ordered                  = 7
-} ompt_mutex_t;
-
-typedef enum ompt_native_mon_flag_t {
-  ompt_native_data_motion_explicit    = 0x01,
-  ompt_native_data_motion_implicit    = 0x02,
-  ompt_native_kernel_invocation       = 0x04,
-  ompt_native_kernel_execution        = 0x08,
-  ompt_native_driver                  = 0x10,
-  ompt_native_runtime                 = 0x20,
-  ompt_native_overhead                = 0x40,
-  ompt_native_idleness                = 0x80
-} ompt_native_mon_flag_t;
-
-typedef enum ompt_task_flag_t {
-  ompt_task_initial                   = 0x00000001,
-  ompt_task_implicit                  = 0x00000002,
-  ompt_task_explicit                  = 0x00000004,
-  ompt_task_target                    = 0x00000008,
-  ompt_task_undeferred                = 0x08000000,
-  ompt_task_untied                    = 0x10000000,
-  ompt_task_final                     = 0x20000000,
-  ompt_task_mergeable                 = 0x40000000,
-  ompt_task_merged                    = 0x80000000
-} ompt_task_flag_t;
-
-typedef enum ompt_task_status_t {
-  ompt_task_complete      = 1,
-  ompt_task_yield         = 2,
-  ompt_task_cancel        = 3,
-  ompt_task_detach        = 4,
-  ompt_task_early_fulfill = 5,
-  ompt_task_late_fulfill  = 6,
-  ompt_task_switch        = 7
-} ompt_task_status_t;
-
-typedef enum ompt_target_t {
-  ompt_target                         = 1,
-  ompt_target_enter_data              = 2,
-  ompt_target_exit_data               = 3,
-  ompt_target_update                  = 4
-} ompt_target_t;
-
-typedef enum ompt_parallel_flag_t {
-  ompt_parallel_invoker_program = 0x00000001,
-  ompt_parallel_invoker_runtime = 0x00000002,
-  ompt_parallel_league          = 0x40000000,
-  ompt_parallel_team            = 0x80000000
-} ompt_parallel_flag_t;
-
-typedef enum ompt_target_map_flag_t {
-  ompt_target_map_flag_to             = 0x01,
-  ompt_target_map_flag_from           = 0x02,
-  ompt_target_map_flag_alloc          = 0x04,
-  ompt_target_map_flag_release        = 0x08,
-  ompt_target_map_flag_delete         = 0x10,
-  ompt_target_map_flag_implicit       = 0x20
-} ompt_target_map_flag_t;
-
-typedef enum ompt_dependence_type_t {
-  ompt_dependence_type_in              = 1,
-  ompt_dependence_type_out             = 2,
-  ompt_dependence_type_inout           = 3,
-  ompt_dependence_type_mutexinoutset   = 4,
-  ompt_dependence_type_source          = 5,
-  ompt_dependence_type_sink            = 6
-} ompt_dependence_type_t;
-
-typedef enum ompt_cancel_flag_t {
-  ompt_cancel_parallel       = 0x01,
-  ompt_cancel_sections       = 0x02,
-  ompt_cancel_loop           = 0x04,
-  ompt_cancel_taskgroup      = 0x08,
-  ompt_cancel_activated      = 0x10,
-  ompt_cancel_detected       = 0x20,
-  ompt_cancel_discarded_task = 0x40
-} ompt_cancel_flag_t;
-
-typedef uint64_t ompt_hwid_t;
-
-typedef uint64_t ompt_wait_id_t;
-
-typedef enum ompt_frame_flag_t {
-  ompt_frame_runtime        = 0x00,
-  ompt_frame_application    = 0x01,
-  ompt_frame_cfa            = 0x10,
-  ompt_frame_framepointer   = 0x20,
-  ompt_frame_stackaddress   = 0x30
-} ompt_frame_flag_t; 
-
-typedef enum ompt_state_t {
-  ompt_state_work_serial                      = 0x000,
-  ompt_state_work_parallel                    = 0x001,
-  ompt_state_work_reduction                   = 0x002,
-
-  ompt_state_wait_barrier                     = 0x010,
-  ompt_state_wait_barrier_implicit_parallel   = 0x011,
-  ompt_state_wait_barrier_implicit_workshare  = 0x012,
-  ompt_state_wait_barrier_implicit            = 0x013,
-  ompt_state_wait_barrier_explicit            = 0x014,
-
-  ompt_state_wait_taskwait                    = 0x020,
-  ompt_state_wait_taskgroup                   = 0x021,
-
-  ompt_state_wait_mutex                       = 0x040,
-  ompt_state_wait_lock                        = 0x041,
-  ompt_state_wait_critical                    = 0x042,
-  ompt_state_wait_atomic                      = 0x043,
-  ompt_state_wait_ordered                     = 0x044,
-
-  ompt_state_wait_target                      = 0x080,
-  ompt_state_wait_target_map                  = 0x081,
-  ompt_state_wait_target_update               = 0x082,
-
-  ompt_state_idle                             = 0x100,
-  ompt_state_overhead                         = 0x101,
-  ompt_state_undefined                        = 0x102
-} ompt_state_t;
-
-typedef uint64_t (*ompt_get_unique_id_t) (void);
-
-typedef uint64_t ompd_size_t;
-
-typedef uint64_t ompd_wait_id_t;
-
-typedef uint64_t ompd_addr_t;
-typedef int64_t  ompd_word_t;
-typedef uint64_t ompd_seg_t;
-
-typedef uint64_t ompd_device_t;
-
-typedef uint64_t ompd_thread_id_t;
-
-typedef enum ompd_scope_t {
-  ompd_scope_global = 1,
-  ompd_scope_address_space = 2,
-  ompd_scope_thread = 3,
-  ompd_scope_parallel = 4,
-  ompd_scope_implicit_task = 5,
-  ompd_scope_task = 6
-} ompd_scope_t;
-
-typedef uint64_t ompd_icv_id_t;
-
-typedef enum ompd_rc_t {
-  ompd_rc_ok = 0,
-  ompd_rc_unavailable = 1,
-  ompd_rc_stale_handle = 2,
-  ompd_rc_bad_input = 3,
-  ompd_rc_error = 4,
-  ompd_rc_unsupported = 5,
-  ompd_rc_needs_state_tracking = 6,
-  ompd_rc_incompatible = 7,
-  ompd_rc_device_read_error = 8,
-  ompd_rc_device_write_error = 9,
-  ompd_rc_nomem = 10,
-} ompd_rc_t;
-
-typedef void (*ompt_interface_fn_t) (void);
-
-typedef ompt_interface_fn_t (*ompt_function_lookup_t) (
-  const char *interface_function_name
-);
-
-typedef union ompt_data_t {
-  uint64_t value;
-  void *ptr;
-} ompt_data_t;
-
-typedef struct ompt_frame_t {
-  ompt_data_t exit_frame;
-  ompt_data_t enter_frame;
-  int exit_frame_flags;
-  int enter_frame_flags;
-} ompt_frame_t;
-
-typedef void (*ompt_callback_t) (void);
-
-typedef void ompt_device_t;
-
-typedef void ompt_buffer_t;
-
-typedef void (*ompt_callback_buffer_request_t) (
-  int device_num,
-  ompt_buffer_t **buffer,
-  size_t *bytes
-);
-
-typedef void (*ompt_callback_buffer_complete_t) (
-  int device_num,
-  ompt_buffer_t *buffer,
-  size_t bytes,
-  ompt_buffer_cursor_t begin,
-  int buffer_owned
-);
-
-typedef void (*ompt_finalize_t) (
-  ompt_data_t *tool_data
-);
-
-typedef int (*ompt_initialize_t) (
-  ompt_function_lookup_t lookup,
-  int initial_device_num,
-  ompt_data_t *tool_data
-);
-
-typedef struct ompt_start_tool_result_t {
-  ompt_initialize_t initialize;
-  ompt_finalize_t finalize;
-  ompt_data_t tool_data;
-} ompt_start_tool_result_t;
-
-typedef struct ompt_record_abstract_t {
-  ompt_record_native_t rclass;
-  const char *type;
-  ompt_device_time_t start_time;
-  ompt_device_time_t end_time;
-  ompt_hwid_t hwid;
-} ompt_record_abstract_t;
-
-typedef struct ompt_dependence_t {
-  ompt_data_t variable;
-  ompt_dependence_type_t dependence_type;
-} ompt_dependence_t;
-
-typedef int (*ompt_enumerate_states_t) (
-  int current_state,
-  int *next_state,
-  const char **next_state_name
-);
-
-typedef int (*ompt_enumerate_mutex_impls_t) (
-  int current_impl,
-  int *next_impl,
-  const char **next_impl_name
-);
-
-typedef ompt_set_result_t (*ompt_set_callback_t) (
-  ompt_callbacks_t event,
-  ompt_callback_t callback
-);
-
-typedef int (*ompt_get_callback_t) (
-  ompt_callbacks_t event,
-  ompt_callback_t *callback
-);
-
-typedef ompt_data_t *(*ompt_get_thread_data_t) (void);
-
-typedef int (*ompt_get_num_procs_t) (void);
-
-typedef int (*ompt_get_num_places_t) (void);
-
-typedef int (*ompt_get_place_proc_ids_t) (
-  int place_num,
-  int ids_size,
-  int *ids
-);
-
-typedef int (*ompt_get_place_num_t) (void);
-
-typedef int (*ompt_get_partition_place_nums_t) (
-  int place_nums_size,
-  int *place_nums
-);
-
-typedef int (*ompt_get_proc_id_t) (void);
-
-typedef int (*ompt_get_state_t) (
-  ompt_wait_id_t *wait_id
-);
-
-typedef int (*ompt_get_parallel_info_t) (
-  int ancestor_level,
-  ompt_data_t **parallel_data,
-  int *team_size
-);
-
-typedef int (*ompt_get_task_info_t) (
-  int ancestor_level,
-  int *flags,
-  ompt_data_t **task_data,
-  ompt_frame_t **task_frame,
-  ompt_data_t **parallel_data,
-  int *thread_num
-);
-
-typedef int (*ompt_get_task_memory_t)(
-  void **addr,
-  size_t *size,
-  int block
-);
-
-typedef int (*ompt_get_target_info_t) (
-  uint64_t *device_num,
-  ompt_id_t *target_id,
-  ompt_id_t *host_op_id
-);
-
-typedef int (*ompt_get_num_devices_t) (void);
-
-typedef void (*ompt_finalize_tool_t) (void);
-
-typedef int (*ompt_get_device_num_procs_t) (
-  ompt_device_t *device
-);
-
-typedef ompt_device_time_t (*ompt_get_device_time_t) (
-  ompt_device_t *device
-);
-
-typedef double (*ompt_translate_time_t) (
-  ompt_device_t *device,
-  ompt_device_time_t time
-);
-
-typedef ompt_set_result_t (*ompt_set_trace_ompt_t) (
-  ompt_device_t *device,
-  unsigned int enable,
-  unsigned int etype
-);
-
-typedef ompt_set_result_t (*ompt_set_trace_native_t) (
-  ompt_device_t *device,
-  int enable,
-  int flags
-);
-
-typedef int (*ompt_start_trace_t) (
-  ompt_device_t *device,
-  ompt_callback_buffer_request_t request,
-  ompt_callback_buffer_complete_t complete
-);
-
-typedef int (*ompt_pause_trace_t) (
-  ompt_device_t *device,
-  int begin_pause
-);
-
-typedef int (*ompt_flush_trace_t) (
-  ompt_device_t *device
-);
-
-typedef int (*ompt_stop_trace_t) (
-  ompt_device_t *device
-);
-
-typedef int (*ompt_advance_buffer_cursor_t) (
-  ompt_device_t *device,
-  ompt_buffer_t *buffer,
-  size_t size,
-  ompt_buffer_cursor_t current,
-  ompt_buffer_cursor_t *next
-);
-
-typedef ompt_record_t (*ompt_get_record_type_t) (
-  ompt_buffer_t *buffer,
-  ompt_buffer_cursor_t current
-);
-
-typedef void *(*ompt_get_record_native_t) (
-  ompt_buffer_t *buffer,
-  ompt_buffer_cursor_t current,
-  ompt_id_t *host_op_id
-);
-
-typedef ompt_record_abstract_t *
-(*ompt_get_record_abstract_t) (
-  void *native_record
-);
-
-typedef void (*ompt_callback_thread_begin_t) (
-  ompt_thread_t thread_type,
-  ompt_data_t *thread_data
-);
-
-typedef struct ompt_record_thread_begin_t {
-  ompt_thread_t thread_type;
-} ompt_record_thread_begin_t;
-
-typedef void (*ompt_callback_thread_end_t) (
-  ompt_data_t *thread_data
-);
-
-typedef void (*ompt_callback_parallel_begin_t) (
-  ompt_data_t *encountering_task_data,
-  const ompt_frame_t *encountering_task_frame,
-  ompt_data_t *parallel_data,
-  unsigned int requested_parallelism,
-  int flags,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_parallel_begin_t {
-  ompt_id_t encountering_task_id;
-  ompt_id_t parallel_id;
-  unsigned int requested_parallelism;
-  int flags;
-  const void *codeptr_ra;
-} ompt_record_parallel_begin_t;
-
-typedef void (*ompt_callback_parallel_end_t) (
-  ompt_data_t *parallel_data,
-  ompt_data_t *encountering_task_data,
-  int flags,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_parallel_end_t {
-  ompt_id_t parallel_id;
-  ompt_id_t encountering_task_id;
-  int flags;
-  const void *codeptr_ra;
-} ompt_record_parallel_end_t;
-
-typedef void (*ompt_callback_work_t) (
-  ompt_work_t wstype,
-  ompt_scope_endpoint_t endpoint,
-  ompt_data_t *parallel_data,
-  ompt_data_t *task_data,
-  uint64_t count,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_work_t {
-  ompt_work_t wstype;
-  ompt_scope_endpoint_t endpoint;
-  ompt_id_t parallel_id;
-  ompt_id_t task_id;
-  uint64_t count;
-  const void *codeptr_ra;
-} ompt_record_work_t;
-
-typedef void (*ompt_callback_dispatch_t) (
-  ompt_data_t *parallel_data,
-  ompt_data_t *task_data,
-  ompt_dispatch_t kind,
-  ompt_data_t instance 
-);
-
-typedef struct ompt_record_dispatch_t {
-  ompt_id_t parallel_id;
-  ompt_id_t task_id;
-  ompt_dispatch_t kind;
-  ompt_data_t instance; 
-} ompt_record_dispatch_t;
-
-typedef void (*ompt_callback_task_create_t) (
-  ompt_data_t *encountering_task_data,
-  const ompt_frame_t *encountering_task_frame,
-  ompt_data_t *new_task_data,
-  int flags,
-  int has_dependences,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_task_create_t {
-  ompt_id_t encountering_task_id;
-  ompt_id_t new_task_id;
-  int flags;
-  int has_dependences;
-  const void *codeptr_ra;
-} ompt_record_task_create_t;
-
-typedef void (*ompt_callback_dependences_t) (
-  ompt_data_t *task_data,
-  const ompt_dependence_t *deps,
-  int ndeps
-);
-
-typedef struct ompt_record_dependences_t {
-  ompt_id_t task_id;
-  ompt_dependence_t dep;
-  int ndeps;
-} ompt_record_dependences_t;
-
-typedef void (*ompt_callback_task_dependence_t) (
-  ompt_data_t *src_task_data,
-  ompt_data_t *sink_task_data
-);
-
-typedef struct ompt_record_task_dependence_t {
-  ompt_id_t src_task_id;
-  ompt_id_t sink_task_id;
-} ompt_record_task_dependence_t;
-
-typedef void (*ompt_callback_task_schedule_t) (
-  ompt_data_t *prior_task_data,
-  ompt_task_status_t prior_task_status,
-  ompt_data_t *next_task_data
-);
-
-typedef struct ompt_record_task_schedule_t {
-  ompt_id_t prior_task_id;
-  ompt_task_status_t prior_task_status;
-  ompt_id_t next_task_id;
-} ompt_record_task_schedule_t;
-
-typedef void (*ompt_callback_implicit_task_t) (
-  ompt_scope_endpoint_t endpoint,
-  ompt_data_t *parallel_data,
-  ompt_data_t *task_data,
-  unsigned int actual_parallelism,
-  unsigned int index,
-  int flags
-);
-
-typedef struct ompt_record_implicit_task_t {
-  ompt_scope_endpoint_t endpoint;
-  ompt_id_t parallel_id;
-  ompt_id_t task_id;
-  unsigned int actual_parallelism;
-  unsigned int index;
-  int flags;
-} ompt_record_implicit_task_t;
-
-typedef void (*ompt_callback_master_t) (
-  ompt_scope_endpoint_t endpoint,
-  ompt_data_t *parallel_data,
-  ompt_data_t *task_data,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_master_t {
-  ompt_scope_endpoint_t endpoint;
-  ompt_id_t parallel_id;
-  ompt_id_t task_id;
-  const void *codeptr_ra;
-} ompt_record_master_t;
-
-typedef void (*ompt_callback_sync_region_t) (
-  ompt_sync_region_t kind,
-  ompt_scope_endpoint_t endpoint,
-  ompt_data_t *parallel_data,
-  ompt_data_t *task_data,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_sync_region_t {
-  ompt_sync_region_t kind;
-  ompt_scope_endpoint_t endpoint;
-  ompt_id_t parallel_id;
-  ompt_id_t task_id;
-  const void *codeptr_ra;
-} ompt_record_sync_region_t;
-
-typedef void (*ompt_callback_mutex_acquire_t) (
-  ompt_mutex_t kind,
-  unsigned int hint,
-  unsigned int impl,
-  ompt_wait_id_t wait_id,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_mutex_acquire_t {
-  ompt_mutex_t kind;
-  unsigned int hint;
-  unsigned int impl;
-  ompt_wait_id_t wait_id;
-  const void *codeptr_ra;
-} ompt_record_mutex_acquire_t;
-
-typedef void (*ompt_callback_mutex_t) (
-  ompt_mutex_t kind,
-  ompt_wait_id_t wait_id,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_mutex_t {
-  ompt_mutex_t kind;
-  ompt_wait_id_t wait_id;
-  const void *codeptr_ra;
-} ompt_record_mutex_t;
-
-typedef void (*ompt_callback_nest_lock_t) (
-  ompt_scope_endpoint_t endpoint,
-  ompt_wait_id_t wait_id,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_nest_lock_t {
-  ompt_scope_endpoint_t endpoint;
-  ompt_wait_id_t wait_id;
-  const void *codeptr_ra;
-} ompt_record_nest_lock_t;
-
-typedef void (*ompt_callback_flush_t) (
-  ompt_data_t *thread_data,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_flush_t {
-  const void *codeptr_ra;
-} ompt_record_flush_t;
-
-typedef void (*ompt_callback_cancel_t) (
-  ompt_data_t *task_data,
-  int flags,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_cancel_t {
-  ompt_id_t task_id;
-  int flags;
-  const void *codeptr_ra;
-} ompt_record_cancel_t;
-
-typedef void (*ompt_callback_device_initialize_t) (
-  int device_num,
-  const char *type,
-  ompt_device_t *device,
-  ompt_function_lookup_t lookup,
-  const char *documentation
-);
-
-typedef void (*ompt_callback_device_finalize_t) (
-  int device_num
-);
-
-typedef void (*ompt_callback_device_load_t) (
-  int device_num,
-  const char *filename,
-  int64_t offset_in_file,
-  void *vma_in_file,
-  size_t bytes,
-  void *host_addr,
-  void *device_addr,
-  uint64_t module_id
-);
-
-typedef void (*ompt_callback_device_unload_t) (
-  int device_num,
-  uint64_t module_id
-);
-
-typedef void (*ompt_callback_target_data_op_t) (
-  ompt_id_t target_id,
-  ompt_id_t host_op_id,
-  ompt_target_data_op_t optype,
-  void *src_addr,
-  int src_device_num,
-  void *dest_addr,
-  int dest_device_num,
-  size_t bytes,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_target_data_op_t {
-  ompt_id_t host_op_id;
-  ompt_target_data_op_t optype;
-  void *src_addr;
-  int src_device_num;
-  void *dest_addr;
-  int dest_device_num;
-  size_t bytes;
-  ompt_device_time_t end_time;
-  const void *codeptr_ra;
-} ompt_record_target_data_op_t;
-
-typedef void (*ompt_callback_target_t) (
-  ompt_target_t kind,
-  ompt_scope_endpoint_t endpoint,
-  int device_num,
-  ompt_data_t *task_data,
-  ompt_id_t target_id,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_target_t {
-  ompt_target_t kind;
-  ompt_scope_endpoint_t endpoint;
-  int device_num;
-  ompt_id_t task_id;
-  ompt_id_t target_id;
-  const void *codeptr_ra;
-} ompt_record_target_t;
-
-typedef void (*ompt_callback_target_map_t) (
-  ompt_id_t target_id,
-  unsigned int nitems,
-  void **host_addr,
-  void **device_addr,
-  size_t *bytes,
-  unsigned int *mapping_flags,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_target_map_t {
-  ompt_id_t target_id;
-  unsigned int nitems;
-  void **host_addr;
-  void **device_addr;
-  size_t *bytes;
-  unsigned int *mapping_flags;
-  const void *codeptr_ra;
-} ompt_record_target_map_t;
-
-typedef void (*ompt_callback_target_submit_t) (
-  ompt_id_t target_id,
-  ompt_id_t host_op_id,
-  unsigned int requested_num_teams
-);
-
-typedef struct ompt_record_target_kernel_t {
-  ompt_id_t host_op_id;
-  unsigned int requested_num_teams;
-  unsigned int granted_num_teams;
-  ompt_device_time_t end_time;
-} ompt_record_target_kernel_t;
-
-typedef int (*ompt_callback_control_tool_t) (
-  uint64_t command,
-  uint64_t modifier,
-  void *arg,
-  const void *codeptr_ra
-);
-
-typedef struct ompt_record_control_tool_t {
-  uint64_t command;
-  uint64_t modifier;
-  const void *codeptr_ra;
-} ompt_record_control_tool_t;
-
-typedef struct ompd_address_t {
-  ompd_seg_t segment;
-  ompd_addr_t address;
-} ompd_address_t;
-
-typedef struct ompd_frame_info_t {
-  ompd_address_t frame_address;
-  ompd_word_t frame_flag;
-} ompd_frame_info_t;
-
-typedef struct _ompd_aspace_handle ompd_address_space_handle_t;
-typedef struct _ompd_thread_handle ompd_thread_handle_t;
-typedef struct _ompd_parallel_handle ompd_parallel_handle_t;
-typedef struct _ompd_task_handle ompd_task_handle_t;
-
-typedef struct _ompd_aspace_cont ompd_address_space_context_t;
-typedef struct _ompd_thread_cont ompd_thread_context_t;
-
-typedef struct ompd_device_type_sizes_t {
-  uint8_t sizeof_char;
-  uint8_t sizeof_short;
-  uint8_t sizeof_int;
-  uint8_t sizeof_long;
-  uint8_t sizeof_long_long;
-  uint8_t sizeof_pointer;
-} ompd_device_type_sizes_t;
-
-typedef struct ompt_record_ompt_t {
-  ompt_callbacks_t type;
-  ompt_device_time_t time;
-  ompt_id_t thread_id;
-  ompt_id_t target_id;
-  union {
-    ompt_record_thread_begin_t thread_begin;
-    ompt_record_parallel_begin_t parallel_begin;
-    ompt_record_parallel_end_t parallel_end;
-    ompt_record_work_t work;
-    ompt_record_dispatch_t dispatch;
-    ompt_record_task_create_t task_create;
-    ompt_record_dependences_t dependences;
-    ompt_record_task_dependence_t task_dependence;
-    ompt_record_task_schedule_t task_schedule;
-    ompt_record_implicit_task_t implicit_task;
-    ompt_record_master_t master;
-    ompt_record_sync_region_t sync_region;
-    ompt_record_mutex_acquire_t mutex_acquire;
-    ompt_record_mutex_t mutex;
-    ompt_record_nest_lock_t nest_lock;
-    ompt_record_flush_t flush;
-    ompt_record_cancel_t cancel;
-    ompt_record_target_t target;
-    ompt_record_target_data_op_t target_data_op;
-    ompt_record_target_map_t target_map;
-    ompt_record_target_kernel_t target_kernel;
-    ompt_record_control_tool_t control_tool;
-  } record;
-} ompt_record_ompt_t;
-
-typedef ompt_record_ompt_t *(*ompt_get_record_ompt_t) (
-  ompt_buffer_t *buffer,
-  ompt_buffer_cursor_t current
-);
-
-#define ompt_id_none 0
-#define ompt_data_none {0}
-#define ompt_time_none 0
-#define ompt_hwid_none 0
-#define ompt_addr_none ~0
-#define ompt_mutex_impl_none 0
-#define ompt_wait_id_none 0
-
-#define ompd_segment_none 0
-
-#endif /* __OMPT__ */
diff --git a/darwin-x86/lib64/libLLVM.dylib b/darwin-x86/lib64/libLLVM.dylib
index ace6735..48bedb3 100755
--- a/darwin-x86/lib64/libLLVM.dylib
+++ b/darwin-x86/lib64/libLLVM.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libbase.dylib b/darwin-x86/lib64/libbase.dylib
index 270c583..c6198e0 100755
--- a/darwin-x86/lib64/libbase.dylib
+++ b/darwin-x86/lib64/libbase.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libc++.1.dylib b/darwin-x86/lib64/libc++.1.dylib
index d430b77..c2540d0 100755
--- a/darwin-x86/lib64/libc++.1.dylib
+++ b/darwin-x86/lib64/libc++.1.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libc++.dylib b/darwin-x86/lib64/libc++.dylib
index ddd4790..bb885dd 100755
--- a/darwin-x86/lib64/libc++.dylib
+++ b/darwin-x86/lib64/libc++.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libc++abi.1.dylib b/darwin-x86/lib64/libc++abi.1.dylib
index ac45654..67a86fe 100755
--- a/darwin-x86/lib64/libc++abi.1.dylib
+++ b/darwin-x86/lib64/libc++abi.1.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libclang_cxx.dylib b/darwin-x86/lib64/libclang_cxx.dylib
index 0591e78..a7842a4 100755
--- a/darwin-x86/lib64/libclang_cxx.dylib
+++ b/darwin-x86/lib64/libclang_cxx.dylib
Binary files differ
diff --git a/darwin-x86/lib64/liblog.dylib b/darwin-x86/lib64/liblog.dylib
index 750d848..d6d83c4 100755
--- a/darwin-x86/lib64/liblog.dylib
+++ b/darwin-x86/lib64/liblog.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libprotobuf-cpp-full.dylib b/darwin-x86/lib64/libprotobuf-cpp-full.dylib
index ed6187e..7586cc2 100755
--- a/darwin-x86/lib64/libprotobuf-cpp-full.dylib
+++ b/darwin-x86/lib64/libprotobuf-cpp-full.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libz-host.dylib b/darwin-x86/lib64/libz-host.dylib
index 9c27db9..d332566 100755
--- a/darwin-x86/lib64/libz-host.dylib
+++ b/darwin-x86/lib64/libz-host.dylib
Binary files differ
diff --git a/darwin-x86/lib64/libziparchive.dylib b/darwin-x86/lib64/libziparchive.dylib
index b3e1981..07627e0 100755
--- a/darwin-x86/lib64/libziparchive.dylib
+++ b/darwin-x86/lib64/libziparchive.dylib
Binary files differ
diff --git a/linux-x86/bin/cxx_extractor b/linux-x86/bin/cxx_extractor
index f8f351a..1be11a0 100755
--- a/linux-x86/bin/cxx_extractor
+++ b/linux-x86/bin/cxx_extractor
Binary files differ
diff --git a/linux-x86/bin/header-abi-diff b/linux-x86/bin/header-abi-diff
index 00e9670..3a9c4c2 100755
--- a/linux-x86/bin/header-abi-diff
+++ b/linux-x86/bin/header-abi-diff
Binary files differ
diff --git a/linux-x86/bin/header-abi-dumper b/linux-x86/bin/header-abi-dumper
index 172e431..00165eb 100755
--- a/linux-x86/bin/header-abi-dumper
+++ b/linux-x86/bin/header-abi-dumper
Binary files differ
diff --git a/linux-x86/bin/header-abi-linker b/linux-x86/bin/header-abi-linker
index c6d26ef..d38fa8c 100755
--- a/linux-x86/bin/header-abi-linker
+++ b/linux-x86/bin/header-abi-linker
Binary files differ
diff --git a/linux-x86/bin/merge-abi-diff b/linux-x86/bin/merge-abi-diff
index 00bc73a..fc26012 100755
--- a/linux-x86/bin/merge-abi-diff
+++ b/linux-x86/bin/merge-abi-diff
Binary files differ
diff --git a/linux-x86/bin/proto_metadata_plugin b/linux-x86/bin/proto_metadata_plugin
new file mode 100755
index 0000000..41dc9c9
--- /dev/null
+++ b/linux-x86/bin/proto_metadata_plugin
Binary files differ
diff --git a/linux-x86/bin/protoc_extractor b/linux-x86/bin/protoc_extractor
index 1a80a7f..f9de13d 100755
--- a/linux-x86/bin/protoc_extractor
+++ b/linux-x86/bin/protoc_extractor
Binary files differ
diff --git a/linux-x86/bin/versioner b/linux-x86/bin/versioner
index 6f8a67a..0bdd48d 100755
--- a/linux-x86/bin/versioner
+++ b/linux-x86/bin/versioner
Binary files differ
diff --git a/linux-x86/clang-headers b/linux-x86/clang-headers
index a5cbaaf..4009f74 120000
--- a/linux-x86/clang-headers
+++ b/linux-x86/clang-headers
@@ -1 +1 @@
-lib64/clang/9.0.7/include
\ No newline at end of file
+lib64/clang/10.0.1/include
\ No newline at end of file
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_builtin_vars.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_builtin_vars.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_builtin_vars.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_builtin_vars.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_cmath.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_cmath.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_cmath.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_cmath.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_complex_builtins.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_complex_builtins.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_complex_builtins.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_complex_builtins.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_device_functions.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_device_functions.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_device_functions.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_device_functions.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
similarity index 98%
copy from darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
copy to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
index 2970d17..b67461a 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
+++ b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_intrinsics.h
@@ -211,7 +211,15 @@
   return __nvvm_vote_ballot_sync(mask, pred);
 }
 
-inline __device__ unsigned int __activemask() { return __nvvm_vote_ballot(1); }
+inline __device__ unsigned int __activemask() {
+#if CUDA_VERSION < 9020
+  return __nvvm_vote_ballot(1);
+#else
+  unsigned int mask;
+  asm volatile("activemask.b32 %0;" : "=r"(mask));
+  return mask;
+#endif
+}
 
 inline __device__ unsigned int __fns(unsigned mask, unsigned base, int offset) {
   return __nvvm_fns(mask, base, offset);
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_libdevice_declares.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_libdevice_declares.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_libdevice_declares.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_libdevice_declares.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_math_forward_declares.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_math_forward_declares.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_math_forward_declares.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_math_forward_declares.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_runtime_wrapper.h b/linux-x86/lib64/clang/10.0.1/include/__clang_cuda_runtime_wrapper.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__clang_cuda_runtime_wrapper.h
rename to linux-x86/lib64/clang/10.0.1/include/__clang_cuda_runtime_wrapper.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__stddef_max_align_t.h b/linux-x86/lib64/clang/10.0.1/include/__stddef_max_align_t.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__stddef_max_align_t.h
rename to linux-x86/lib64/clang/10.0.1/include/__stddef_max_align_t.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__wmmintrin_aes.h b/linux-x86/lib64/clang/10.0.1/include/__wmmintrin_aes.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__wmmintrin_aes.h
rename to linux-x86/lib64/clang/10.0.1/include/__wmmintrin_aes.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/__wmmintrin_pclmul.h b/linux-x86/lib64/clang/10.0.1/include/__wmmintrin_pclmul.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/__wmmintrin_pclmul.h
rename to linux-x86/lib64/clang/10.0.1/include/__wmmintrin_pclmul.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/adxintrin.h b/linux-x86/lib64/clang/10.0.1/include/adxintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/adxintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/adxintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/altivec.h b/linux-x86/lib64/clang/10.0.1/include/altivec.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/altivec.h
rename to linux-x86/lib64/clang/10.0.1/include/altivec.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/ammintrin.h b/linux-x86/lib64/clang/10.0.1/include/ammintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/ammintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/ammintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/arm64intr.h b/linux-x86/lib64/clang/10.0.1/include/arm64intr.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/arm64intr.h
rename to linux-x86/lib64/clang/10.0.1/include/arm64intr.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm_acle.h b/linux-x86/lib64/clang/10.0.1/include/arm_acle.h
similarity index 94%
copy from darwin-x86/lib64/clang/9.0.7/include/arm_acle.h
copy to linux-x86/lib64/clang/10.0.1/include/arm_acle.h
index 08d65fa..0510e6f 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/arm_acle.h
+++ b/linux-x86/lib64/clang/10.0.1/include/arm_acle.h
@@ -597,6 +597,14 @@
 }
 #endif
 
+/* Armv8.3-A Javascript conversion intrinsic */
+#if __ARM_64BIT_STATE && defined(__ARM_FEATURE_JCVT)
+static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
+__jcvt(double __a) {
+  return __builtin_arm_jcvt(__a);
+}
+#endif
+
 /* 10.1 Special register intrinsics */
 #define __arm_rsr(sysreg) __builtin_arm_rsr(sysreg)
 #define __arm_rsr64(sysreg) __builtin_arm_rsr64(sysreg)
@@ -605,7 +613,7 @@
 #define __arm_wsr64(sysreg, v) __builtin_arm_wsr64(sysreg, v)
 #define __arm_wsrp(sysreg, v) __builtin_arm_wsrp(sysreg, v)
 
-// Memory Tagging Extensions (MTE) Intrinsics
+/* Memory Tagging Extensions (MTE) Intrinsics */
 #if __ARM_FEATURE_MEMORY_TAGGING
 #define __arm_mte_create_random_tag(__ptr, __mask)  __builtin_arm_irg(__ptr, __mask)
 #define __arm_mte_increment_tag(__ptr, __tag_offset)  __builtin_arm_addg(__ptr, __tag_offset)
@@ -615,6 +623,28 @@
 #define __arm_mte_ptrdiff(__ptra, __ptrb) __builtin_arm_subp(__ptra, __ptrb)
 #endif
 
+/* Transactional Memory Extension (TME) Intrinsics */
+#if __ARM_FEATURE_TME
+
+#define _TMFAILURE_REASON  0x00007fffu
+#define _TMFAILURE_RTRY    0x00008000u
+#define _TMFAILURE_CNCL    0x00010000u
+#define _TMFAILURE_MEM     0x00020000u
+#define _TMFAILURE_IMP     0x00040000u
+#define _TMFAILURE_ERR     0x00080000u
+#define _TMFAILURE_SIZE    0x00100000u
+#define _TMFAILURE_NEST    0x00200000u
+#define _TMFAILURE_DBG     0x00400000u
+#define _TMFAILURE_INT     0x00800000u
+#define _TMFAILURE_TRIVIAL 0x01000000u
+
+#define __tstart()        __builtin_arm_tstart()
+#define __tcommit()       __builtin_arm_tcommit()
+#define __tcancel(__arg)  __builtin_arm_tcancel(__arg)
+#define __ttest()         __builtin_arm_ttest()
+
+#endif /* __ARM_FEATURE_TME */
+
 #if defined(__cplusplus)
 }
 #endif
diff --git a/linux-x86/lib64/clang/10.0.1/include/arm_fp16.h b/linux-x86/lib64/clang/10.0.1/include/arm_fp16.h
new file mode 100644
index 0000000..1c04f2a
--- /dev/null
+++ b/linux-x86/lib64/clang/10.0.1/include/arm_fp16.h
@@ -0,0 +1,590 @@
+/*===---- arm_fp16.h - ARM FP16 intrinsics ---------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#ifndef __ARM_FP16_H
+#define __ARM_FP16_H
+
+#include <stdint.h>
+
+typedef __fp16 float16_t;
+#define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
+
+#if defined(__ARM_FEATURE_FP16_SCALAR_ARITHMETIC) && defined(__aarch64__)
+#define vabdh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vabsh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
+  __ret; \
+})
+#define vaddh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcageh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcagth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcaleh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcalth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vceqh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vceqzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
+  __ret; \
+})
+#define vcgeh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcgezh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
+  __ret; \
+})
+#define vcgth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcgtzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
+  __ret; \
+})
+#define vcleh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vclezh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
+  __ret; \
+})
+#define vclth_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
+  __ret; \
+})
+#define vcltzh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
+  __ret; \
+})
+#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvth_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvth_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvth_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtah_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
+  __ret; \
+})
+__ai float16_t vcvth_f16_u32(uint32_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_u64(uint64_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_u16(uint16_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s32(int32_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s64(int64_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
+  return __ret;
+}
+__ai float16_t vcvth_f16_s16(int16_t __p0) {
+  float16_t __ret;
+  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
+  return __ret;
+}
+#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
+  __ret; \
+})
+#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
+  __ret; \
+})
+#define vcvtmh_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtmh_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtnh_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_s64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u16_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u32_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
+  __ret; \
+})
+#define vcvtph_u64_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
+  __ret; \
+})
+#define vdivh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __s2 = __p2; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
+  __ret; \
+})
+#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __s2 = __p2; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
+  __ret; \
+})
+#define vmaxh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vminh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vminnmh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmulh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vmulxh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vnegh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
+  __ret; \
+})
+#define vrecpeh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
+  __ret; \
+})
+#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vrecpxh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
+  __ret; \
+})
+#define vrndh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
+  __ret; \
+})
+#define vrndah_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
+  __ret; \
+})
+#define vrndih_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
+  __ret; \
+})
+#define vrndmh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
+  __ret; \
+})
+#define vrndnh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
+  __ret; \
+})
+#define vrndph_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
+  __ret; \
+})
+#define vrndxh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
+  __ret; \
+})
+#define vrsqrteh_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
+  __ret; \
+})
+#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
+  __ret; \
+})
+#define vsqrth_f16(__p0) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
+  __ret; \
+})
+#define vsubh_f16(__p0, __p1) __extension__ ({ \
+  float16_t __s0 = __p0; \
+  float16_t __s1 = __p1; \
+  float16_t __ret; \
+  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
+  __ret; \
+})
+#endif
+
+#undef __ai
+
+#endif /* __ARM_FP16_H */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/arm_neon.h b/linux-x86/lib64/clang/10.0.1/include/arm_neon.h
similarity index 86%
copy from darwin-x86/lib64/clang/9.0.7/include/arm_neon.h
copy to linux-x86/lib64/clang/10.0.1/include/arm_neon.h
index 694bdfc..6b55c4b 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/arm_neon.h
+++ b/linux-x86/lib64/clang/10.0.1/include/arm_neon.h
@@ -1042,20 +1042,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1124,20 +1115,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -1457,20 +1439,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 & __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1522,20 +1495,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 & __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -1723,20 +1687,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 & ~__p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -1788,20 +1743,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 & ~__p1;
   return __ret;
 }
-#else
-__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -2089,20 +2035,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
   uint16x4_t __ret;
@@ -2175,20 +2112,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
   int16x4_t __ret;
@@ -4191,174 +4119,66 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vcreate_p8(uint64_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vcreate_p8(uint64_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vcreate_p16(uint64_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vcreate_p16(uint64_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vcreate_u8(uint64_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vcreate_u8(uint64_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcreate_u32(uint64_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vcreate_u32(uint64_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcreate_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vcreate_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcreate_u16(uint64_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vcreate_u16(uint64_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vcreate_s8(uint64_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vcreate_s8(uint64_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcreate_f32(uint64_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vcreate_f32(uint64_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vcreate_f16(uint64_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vcreate_f16(uint64_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vcreate_s32(uint64_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vcreate_s32(uint64_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcreate_s64(uint64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vcreate_s64(uint64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vcreate_s16(uint64_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vcreate_s16(uint64_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
@@ -4899,22 +4719,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -4987,22 +4797,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -5263,20 +5063,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vdup_n_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vdup_n_u16(uint16_t __p0) {
   uint16x4_t __ret;
@@ -5354,20 +5145,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vdup_n_s64(int64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai int64x1_t vdup_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vdup_n_s16(int16_t __p0) {
   int16x4_t __ret;
@@ -5553,20 +5335,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 ^ __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -5618,20 +5391,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 ^ __p1;
   return __ret;
 }
-#else
-__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -5964,7 +5728,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -5972,16 +5735,6 @@
   __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vext_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -6066,7 +5819,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -6074,16 +5826,6 @@
   __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vext_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -6685,28 +6427,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vget_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -6799,28 +6525,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vget_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -7667,20 +7377,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64(__p0) __extension__ ({ \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16(__p0) __extension__ ({ \
   uint16x4_t __ret; \
@@ -7741,20 +7442,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64(__p0) __extension__ ({ \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16(__p0) __extension__ ({ \
   int16x4_t __ret; \
@@ -7995,20 +7687,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_u64(__p0) __extension__ ({ \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
   __ret; \
 })
-#else
-#define vld1_dup_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_dup_u16(__p0) __extension__ ({ \
   uint16x4_t __ret; \
@@ -8069,20 +7752,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_s64(__p0) __extension__ ({ \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
   __ret; \
 })
-#else
-#define vld1_dup_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_dup_s16(__p0) __extension__ ({ \
   int16x4_t __ret; \
@@ -8368,22 +8042,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -8456,22 +8120,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -8745,20 +8399,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x2(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x2(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x2(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -8827,20 +8472,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x2(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x2(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x2(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -9128,20 +8764,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x3(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x3(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x3(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -9214,20 +8841,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x3(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x3(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x3(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -9531,20 +9149,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_u64_x4(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld1_u64_x4(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_u16_x4(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -9621,20 +9230,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_s64_x4(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld1_s64_x4(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1_s16_x4(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -9875,20 +9475,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_u64(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld2_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_u16(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -9957,20 +9548,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_s64(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld2_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_s16(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -10243,20 +9825,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_u64(__p0) __extension__ ({ \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld2_dup_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_dup_u16(__p0) __extension__ ({ \
   uint16x4x2_t __ret; \
@@ -10325,20 +9898,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_s64(__p0) __extension__ ({ \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld2_dup_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2_dup_s16(__p0) __extension__ ({ \
   int16x4x2_t __ret; \
@@ -10920,20 +10484,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_u64(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld3_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_u16(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -11006,20 +10561,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_s64(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld3_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_s16(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -11308,20 +10854,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_u64(__p0) __extension__ ({ \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld3_dup_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_dup_u16(__p0) __extension__ ({ \
   uint16x4x3_t __ret; \
@@ -11394,20 +10931,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_s64(__p0) __extension__ ({ \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld3_dup_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3_dup_s16(__p0) __extension__ ({ \
   int16x4x3_t __ret; \
@@ -12033,20 +11561,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_u64(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld4_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_u16(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -12123,20 +11642,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_s64(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld4_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_s16(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -12441,20 +11951,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_u64(__p0) __extension__ ({ \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
   __ret; \
 })
-#else
-#define vld4_dup_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_dup_u16(__p0) __extension__ ({ \
   uint16x4x4_t __ret; \
@@ -12531,20 +12032,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_s64(__p0) __extension__ ({ \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
   __ret; \
 })
-#else
-#define vld4_dup_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4_dup_s16(__p0) __extension__ ({ \
   int16x4x4_t __ret; \
@@ -14996,20 +14488,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vmov_n_u64(uint64_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vmov_n_u16(uint16_t __p0) {
   uint16x4_t __ret;
@@ -15087,20 +14570,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vmov_n_s64(int64_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai int64x1_t vmov_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vmov_n_s16(int16_t __p0) {
   int16x4_t __ret;
@@ -16854,20 +16328,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 | ~__p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -16919,20 +16384,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 | ~__p1;
   return __ret;
 }
-#else
-__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -17120,20 +16576,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 | __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -17185,20 +16632,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 | __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -18239,20 +17677,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -18309,20 +17738,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -19676,20 +19096,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -19741,20 +19152,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -20158,20 +19560,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -20223,20 +19616,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -20434,22 +19818,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshl_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshl_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -20504,22 +19878,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshl_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshl_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -20646,22 +20010,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -21076,20 +20430,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -21146,20 +20491,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -22362,20 +21698,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -22427,20 +21754,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -22638,22 +21956,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrshr_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrshr_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -22708,22 +22016,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrshr_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrshr_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -23194,7 +22492,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -23202,16 +22499,6 @@
   __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -23275,7 +22562,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -23283,16 +22569,6 @@
   __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -23851,7 +23127,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -23859,23 +23134,6 @@
   __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16_t __s0 = __p0; \
@@ -23984,7 +23242,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -23992,23 +23249,6 @@
   __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16_t __s0 = __p0; \
@@ -24206,20 +23446,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -24271,20 +23502,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -24482,22 +23704,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshl_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshl_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -24552,22 +23764,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshl_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshl_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -24910,22 +24112,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshr_n_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshr_n_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -24980,22 +24172,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vshr_n_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vshr_n_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -25452,7 +24634,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -25460,16 +24641,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -25533,7 +24704,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -25541,16 +24711,6 @@
   __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -25782,7 +24942,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -25790,16 +24949,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -25863,7 +25012,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -25871,16 +25019,6 @@
   __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -26196,7 +25334,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64x1_t __s1 = __p1; \
@@ -26204,16 +25341,6 @@
   __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
   __ret; \
 })
-#else
-#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -26277,7 +25404,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64x1_t __s1 = __p1; \
@@ -26285,16 +25411,6 @@
   __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
   __ret; \
 })
-#else
-#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -26511,18 +25627,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
 })
-#else
-#define vst1_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -26575,18 +25683,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
 })
-#else
-#define vst1_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -26795,18 +25895,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
 })
-#else
-#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
   uint16x4_t __s1 = __p1; \
@@ -26859,18 +25951,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
 })
-#else
-#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
   int16x4_t __s1 = __p1; \
@@ -27109,18 +26193,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x2(__p0, __p1) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
 })
-#else
-#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x2(__p0, __p1) __extension__ ({ \
   uint16x4x2_t __s1 = __p1; \
@@ -27181,18 +26257,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x2(__p0, __p1) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
 })
-#else
-#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x2(__p0, __p1) __extension__ ({ \
   int16x4x2_t __s1 = __p1; \
@@ -27448,18 +26516,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x3(__p0, __p1) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
 })
-#else
-#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x3(__p0, __p1) __extension__ ({ \
   uint16x4x3_t __s1 = __p1; \
@@ -27524,18 +26584,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x3(__p0, __p1) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
 })
-#else
-#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x3(__p0, __p1) __extension__ ({ \
   int16x4x3_t __s1 = __p1; \
@@ -27807,18 +26859,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_u64_x4(__p0, __p1) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
 })
-#else
-#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_u16_x4(__p0, __p1) __extension__ ({ \
   uint16x4x4_t __s1 = __p1; \
@@ -27887,18 +26931,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_s64_x4(__p0, __p1) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
 })
-#else
-#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1_s16_x4(__p0, __p1) __extension__ ({ \
   int16x4x4_t __s1 = __p1; \
@@ -28111,18 +27147,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_u64(__p0, __p1) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
 })
-#else
-#define vst2_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2_u16(__p0, __p1) __extension__ ({ \
   uint16x4x2_t __s1 = __p1; \
@@ -28183,18 +27211,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_s64(__p0, __p1) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
 })
-#else
-#define vst2_s64(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2_s16(__p0, __p1) __extension__ ({ \
   int16x4x2_t __s1 = __p1; \
@@ -28643,18 +27663,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_u64(__p0, __p1) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
 })
-#else
-#define vst3_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3_u16(__p0, __p1) __extension__ ({ \
   uint16x4x3_t __s1 = __p1; \
@@ -28719,18 +27731,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_s64(__p0, __p1) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
 })
-#else
-#define vst3_s64(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3_s16(__p0, __p1) __extension__ ({ \
   int16x4x3_t __s1 = __p1; \
@@ -29208,18 +28212,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_u64(__p0, __p1) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
 })
-#else
-#define vst4_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4_u16(__p0, __p1) __extension__ ({ \
   uint16x4x4_t __s1 = __p1; \
@@ -29288,18 +28284,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_s64(__p0, __p1) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
 })
-#else
-#define vst4_s64(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4_s16(__p0, __p1) __extension__ ({ \
   int16x4x4_t __s1 = __p1; \
@@ -29759,20 +28747,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
   uint16x4_t __ret;
@@ -29841,20 +28820,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
   int64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
   int16x4_t __ret;
@@ -32085,3702 +31055,1326 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #endif
 #if (__ARM_FP & 2)
 #ifdef __LITTLE_ENDIAN__
@@ -37072,20 +33666,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vsha1h_u32(uint32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vsha1h_u32(uint32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
   uint32x4_t __ret;
@@ -37388,20 +33973,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrndns_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrndns_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vrndpq_f32(float32x4_t __p0) {
   float32x4_t __ret;
@@ -37818,20 +34394,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37848,20 +34415,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37878,20 +34436,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37908,20 +34457,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37938,20 +34478,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -37968,20 +34499,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -37998,20 +34520,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -38028,5508 +34541,1971 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
   poly8x8_t __ret;
   __ret = (poly8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
   poly16x4_t __ret;
   __ret = (poly16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
   poly8x16_t __ret;
   __ret = (poly8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
   poly128_t __ret;
   __ret = (poly128_t)(__p0);
   return __ret;
 }
-#else
-__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
   poly64x2_t __ret;
   __ret = (poly64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
   poly16x8_t __ret;
   __ret = (poly16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
   uint8x16_t __ret;
   __ret = (uint8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
   uint32x4_t __ret;
   __ret = (uint32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
   uint64x2_t __ret;
   __ret = (uint64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
   uint16x8_t __ret;
   __ret = (uint16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
   int8x16_t __ret;
   __ret = (int8x16_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
   float32x4_t __ret;
   __ret = (float32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
   float16x8_t __ret;
   __ret = (float16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
   int32x4_t __ret;
   __ret = (int32x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
   int64x2_t __ret;
   __ret = (int64x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
   int16x8_t __ret;
   __ret = (int16x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
   uint8x8_t __ret;
   __ret = (uint8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
   uint32x2_t __ret;
   __ret = (uint32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
   uint16x4_t __ret;
   __ret = (uint16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
   int8x8_t __ret;
   __ret = (int8x8_t)(__p0);
   return __ret;
 }
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
   float32x2_t __ret;
   __ret = (float32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
   float16x4_t __ret;
   __ret = (float16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
   int32x2_t __ret;
   __ret = (int32x2_t)(__p0);
   return __ret;
 }
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
   int16x4_t __ret;
   __ret = (int16x4_t)(__p0);
   return __ret;
 }
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
 #endif
 #if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
 #ifdef __LITTLE_ENDIAN__
@@ -43548,20 +36524,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrnd_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrnd_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndaq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43578,20 +36545,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrnda_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrnda_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndiq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43608,20 +36566,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndi_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndi_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndmq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43638,20 +36587,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndm_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndm_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndnq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43668,20 +36608,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndn_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndn_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndpq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43698,20 +36629,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndp_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndp_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrndxq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -43728,20 +36650,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrndx_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrndx_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #endif
 #if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
 #ifdef __LITTLE_ENDIAN__
@@ -43761,20 +36674,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -43792,20 +36696,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #endif
 #if defined(__ARM_FEATURE_DOTPROD)
 #ifdef __LITTLE_ENDIAN__
@@ -48073,48 +40968,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vabsq_f64(float64x2_t __p0) {
   float64x2_t __ret;
@@ -48147,48 +41015,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vabs_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vabs_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vabs_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vabsd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -48206,48 +41047,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 + __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -48791,20 +41605,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
   return __ret;
 }
-#else
-__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
   poly64x2_t __ret;
@@ -48841,20 +41646,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48872,48 +41668,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48931,48 +41700,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -48990,48 +41732,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
   uint64x2_t __ret;
@@ -49049,62 +41764,26 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
   uint64x2_t __ret;
@@ -49173,104 +41852,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 == __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
   uint8x8_t __ret;
@@ -49287,20 +41903,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
   uint16x4_t __ret;
@@ -49557,20 +42164,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
   uint16x4_t __ret;
@@ -49603,20 +42201,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vceqz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -49649,20 +42238,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vceqz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vceqz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -49679,62 +42259,26 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqzd_u64(uint64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vceqzd_u64(uint64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vceqzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vceqzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vceqzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vceqzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vceqzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vceqzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -49786,104 +42330,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 >= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -49996,20 +42477,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgez_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcgez_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50042,20 +42514,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgez_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcgez_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50072,48 +42535,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgezd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcgezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgezd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcgezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgezs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcgezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50165,104 +42601,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 > __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -50375,20 +42748,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50421,20 +42785,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50451,48 +42806,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcgtzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcgtzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcgtzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcgtzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcgtzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcgtzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50544,104 +42872,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 <= __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vclezq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -50754,20 +43019,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclez_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vclez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vclez_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -50800,20 +43056,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclez_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vclez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vclez_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -50830,48 +43077,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vclezd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vclezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vclezd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vclezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vclezs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vclezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
   uint64x2_t __ret;
@@ -50923,104 +43143,41 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t)(__p0 < __p1);
   return __ret;
 }
-#else
-__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
   uint8x16_t __ret;
@@ -51133,20 +43290,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcltz_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint32x2_t vcltz_f32(float32x2_t __p0) {
   uint32x2_t __ret;
@@ -51179,20 +43327,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcltz_s64(int64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vcltz_s16(int16x4_t __p0) {
   uint16x4_t __ret;
@@ -51209,48 +43348,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcltzd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcltzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcltzd_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcltzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcltzs_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcltzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
   poly64x2_t __ret;
@@ -51379,7 +43491,7 @@
   uint64x1_t __s2_21 = __p2_21; \
   uint64x2_t __rev0_21;  __rev0_21 = __builtin_shufflevector(__s0_21, __s0_21, 1, 0); \
   uint64x2_t __ret_21; \
-  __ret_21 = __noswap_vsetq_lane_u64(__noswap_vget_lane_u64(__s2_21, __p3_21), __rev0_21, __p1_21); \
+  __ret_21 = __noswap_vsetq_lane_u64(vget_lane_u64(__s2_21, __p3_21), __rev0_21, __p1_21); \
   __ret_21 = __builtin_shufflevector(__ret_21, __ret_21, 1, 0); \
   __ret_21; \
 })
@@ -51483,7 +43595,7 @@
   int64x1_t __s2_31 = __p2_31; \
   int64x2_t __rev0_31;  __rev0_31 = __builtin_shufflevector(__s0_31, __s0_31, 1, 0); \
   int64x2_t __ret_31; \
-  __ret_31 = __noswap_vsetq_lane_s64(__noswap_vget_lane_s64(__s2_31, __p3_31), __rev0_31, __p1_31); \
+  __ret_31 = __noswap_vsetq_lane_s64(vget_lane_s64(__s2_31, __p3_31), __rev0_31, __p1_31); \
   __ret_31 = __builtin_shufflevector(__ret_31, __ret_31, 1, 0); \
   __ret_31; \
 })
@@ -51594,7 +43706,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcopy_lane_u64(__p0_42, __p1_42, __p2_42, __p3_42) __extension__ ({ \
   uint64x1_t __s0_42 = __p0_42; \
   uint64x1_t __s2_42 = __p2_42; \
@@ -51602,653 +43713,596 @@
   __ret_42 = vset_lane_u64(vget_lane_u64(__s2_42, __p3_42), __s0_42, __p1_42); \
   __ret_42; \
 })
-#else
-#define vcopy_lane_u64(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
-  uint64x1_t __s0_43 = __p0_43; \
-  uint64x1_t __s2_43 = __p2_43; \
-  uint64x1_t __ret_43; \
-  __ret_43 = __noswap_vset_lane_u64(__noswap_vget_lane_u64(__s2_43, __p3_43), __s0_43, __p1_43); \
+#ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_u16(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
+  uint16x4_t __s0_43 = __p0_43; \
+  uint16x4_t __s2_43 = __p2_43; \
+  uint16x4_t __ret_43; \
+  __ret_43 = vset_lane_u16(vget_lane_u16(__s2_43, __p3_43), __s0_43, __p1_43); \
   __ret_43; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
+#else
 #define vcopy_lane_u16(__p0_44, __p1_44, __p2_44, __p3_44) __extension__ ({ \
   uint16x4_t __s0_44 = __p0_44; \
   uint16x4_t __s2_44 = __p2_44; \
+  uint16x4_t __rev0_44;  __rev0_44 = __builtin_shufflevector(__s0_44, __s0_44, 3, 2, 1, 0); \
+  uint16x4_t __rev2_44;  __rev2_44 = __builtin_shufflevector(__s2_44, __s2_44, 3, 2, 1, 0); \
   uint16x4_t __ret_44; \
-  __ret_44 = vset_lane_u16(vget_lane_u16(__s2_44, __p3_44), __s0_44, __p1_44); \
+  __ret_44 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_44, __p3_44), __rev0_44, __p1_44); \
+  __ret_44 = __builtin_shufflevector(__ret_44, __ret_44, 3, 2, 1, 0); \
   __ret_44; \
 })
-#else
-#define vcopy_lane_u16(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
-  uint16x4_t __s0_45 = __p0_45; \
-  uint16x4_t __s2_45 = __p2_45; \
-  uint16x4_t __rev0_45;  __rev0_45 = __builtin_shufflevector(__s0_45, __s0_45, 3, 2, 1, 0); \
-  uint16x4_t __rev2_45;  __rev2_45 = __builtin_shufflevector(__s2_45, __s2_45, 3, 2, 1, 0); \
-  uint16x4_t __ret_45; \
-  __ret_45 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_45, __p3_45), __rev0_45, __p1_45); \
-  __ret_45 = __builtin_shufflevector(__ret_45, __ret_45, 3, 2, 1, 0); \
-  __ret_45; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s8(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
+  int8x8_t __s0_45 = __p0_45; \
+  int8x8_t __s2_45 = __p2_45; \
+  int8x8_t __ret_45; \
+  __ret_45 = vset_lane_s8(vget_lane_s8(__s2_45, __p3_45), __s0_45, __p1_45); \
+  __ret_45; \
+})
+#else
 #define vcopy_lane_s8(__p0_46, __p1_46, __p2_46, __p3_46) __extension__ ({ \
   int8x8_t __s0_46 = __p0_46; \
   int8x8_t __s2_46 = __p2_46; \
+  int8x8_t __rev0_46;  __rev0_46 = __builtin_shufflevector(__s0_46, __s0_46, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __rev2_46;  __rev2_46 = __builtin_shufflevector(__s2_46, __s2_46, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x8_t __ret_46; \
-  __ret_46 = vset_lane_s8(vget_lane_s8(__s2_46, __p3_46), __s0_46, __p1_46); \
+  __ret_46 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_46, __p3_46), __rev0_46, __p1_46); \
+  __ret_46 = __builtin_shufflevector(__ret_46, __ret_46, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_46; \
 })
-#else
-#define vcopy_lane_s8(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
-  int8x8_t __s0_47 = __p0_47; \
-  int8x8_t __s2_47 = __p2_47; \
-  int8x8_t __rev0_47;  __rev0_47 = __builtin_shufflevector(__s0_47, __s0_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2_47;  __rev2_47 = __builtin_shufflevector(__s2_47, __s2_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_47; \
-  __ret_47 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_47, __p3_47), __rev0_47, __p1_47); \
-  __ret_47 = __builtin_shufflevector(__ret_47, __ret_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_47; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_f32(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
+  float32x2_t __s0_47 = __p0_47; \
+  float32x2_t __s2_47 = __p2_47; \
+  float32x2_t __ret_47; \
+  __ret_47 = vset_lane_f32(vget_lane_f32(__s2_47, __p3_47), __s0_47, __p1_47); \
+  __ret_47; \
+})
+#else
 #define vcopy_lane_f32(__p0_48, __p1_48, __p2_48, __p3_48) __extension__ ({ \
   float32x2_t __s0_48 = __p0_48; \
   float32x2_t __s2_48 = __p2_48; \
+  float32x2_t __rev0_48;  __rev0_48 = __builtin_shufflevector(__s0_48, __s0_48, 1, 0); \
+  float32x2_t __rev2_48;  __rev2_48 = __builtin_shufflevector(__s2_48, __s2_48, 1, 0); \
   float32x2_t __ret_48; \
-  __ret_48 = vset_lane_f32(vget_lane_f32(__s2_48, __p3_48), __s0_48, __p1_48); \
+  __ret_48 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_48, __p3_48), __rev0_48, __p1_48); \
+  __ret_48 = __builtin_shufflevector(__ret_48, __ret_48, 1, 0); \
   __ret_48; \
 })
-#else
-#define vcopy_lane_f32(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
-  float32x2_t __s0_49 = __p0_49; \
-  float32x2_t __s2_49 = __p2_49; \
-  float32x2_t __rev0_49;  __rev0_49 = __builtin_shufflevector(__s0_49, __s0_49, 1, 0); \
-  float32x2_t __rev2_49;  __rev2_49 = __builtin_shufflevector(__s2_49, __s2_49, 1, 0); \
-  float32x2_t __ret_49; \
-  __ret_49 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_49, __p3_49), __rev0_49, __p1_49); \
-  __ret_49 = __builtin_shufflevector(__ret_49, __ret_49, 1, 0); \
-  __ret_49; \
-})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
+#define vcopy_lane_s32(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
+  int32x2_t __s0_49 = __p0_49; \
+  int32x2_t __s2_49 = __p2_49; \
+  int32x2_t __ret_49; \
+  __ret_49 = vset_lane_s32(vget_lane_s32(__s2_49, __p3_49), __s0_49, __p1_49); \
+  __ret_49; \
+})
+#else
 #define vcopy_lane_s32(__p0_50, __p1_50, __p2_50, __p3_50) __extension__ ({ \
   int32x2_t __s0_50 = __p0_50; \
   int32x2_t __s2_50 = __p2_50; \
+  int32x2_t __rev0_50;  __rev0_50 = __builtin_shufflevector(__s0_50, __s0_50, 1, 0); \
+  int32x2_t __rev2_50;  __rev2_50 = __builtin_shufflevector(__s2_50, __s2_50, 1, 0); \
   int32x2_t __ret_50; \
-  __ret_50 = vset_lane_s32(vget_lane_s32(__s2_50, __p3_50), __s0_50, __p1_50); \
+  __ret_50 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_50, __p3_50), __rev0_50, __p1_50); \
+  __ret_50 = __builtin_shufflevector(__ret_50, __ret_50, 1, 0); \
   __ret_50; \
 })
-#else
-#define vcopy_lane_s32(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
-  int32x2_t __s0_51 = __p0_51; \
-  int32x2_t __s2_51 = __p2_51; \
-  int32x2_t __rev0_51;  __rev0_51 = __builtin_shufflevector(__s0_51, __s0_51, 1, 0); \
-  int32x2_t __rev2_51;  __rev2_51 = __builtin_shufflevector(__s2_51, __s2_51, 1, 0); \
-  int32x2_t __ret_51; \
-  __ret_51 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_51, __p3_51), __rev0_51, __p1_51); \
-  __ret_51 = __builtin_shufflevector(__ret_51, __ret_51, 1, 0); \
-  __ret_51; \
-})
 #endif
 
+#define vcopy_lane_s64(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
+  int64x1_t __s0_51 = __p0_51; \
+  int64x1_t __s2_51 = __p2_51; \
+  int64x1_t __ret_51; \
+  __ret_51 = vset_lane_s64(vget_lane_s64(__s2_51, __p3_51), __s0_51, __p1_51); \
+  __ret_51; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s64(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
-  int64x1_t __s0_52 = __p0_52; \
-  int64x1_t __s2_52 = __p2_52; \
-  int64x1_t __ret_52; \
-  __ret_52 = vset_lane_s64(vget_lane_s64(__s2_52, __p3_52), __s0_52, __p1_52); \
+#define vcopy_lane_s16(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
+  int16x4_t __s0_52 = __p0_52; \
+  int16x4_t __s2_52 = __p2_52; \
+  int16x4_t __ret_52; \
+  __ret_52 = vset_lane_s16(vget_lane_s16(__s2_52, __p3_52), __s0_52, __p1_52); \
   __ret_52; \
 })
 #else
-#define vcopy_lane_s64(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
-  int64x1_t __s0_53 = __p0_53; \
-  int64x1_t __s2_53 = __p2_53; \
-  int64x1_t __ret_53; \
-  __ret_53 = __noswap_vset_lane_s64(__noswap_vget_lane_s64(__s2_53, __p3_53), __s0_53, __p1_53); \
+#define vcopy_lane_s16(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
+  int16x4_t __s0_53 = __p0_53; \
+  int16x4_t __s2_53 = __p2_53; \
+  int16x4_t __rev0_53;  __rev0_53 = __builtin_shufflevector(__s0_53, __s0_53, 3, 2, 1, 0); \
+  int16x4_t __rev2_53;  __rev2_53 = __builtin_shufflevector(__s2_53, __s2_53, 3, 2, 1, 0); \
+  int16x4_t __ret_53; \
+  __ret_53 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_53, __p3_53), __rev0_53, __p1_53); \
+  __ret_53 = __builtin_shufflevector(__ret_53, __ret_53, 3, 2, 1, 0); \
   __ret_53; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s16(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
-  int16x4_t __s0_54 = __p0_54; \
-  int16x4_t __s2_54 = __p2_54; \
-  int16x4_t __ret_54; \
-  __ret_54 = vset_lane_s16(vget_lane_s16(__s2_54, __p3_54), __s0_54, __p1_54); \
+#define vcopyq_laneq_p8(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
+  poly8x16_t __s0_54 = __p0_54; \
+  poly8x16_t __s2_54 = __p2_54; \
+  poly8x16_t __ret_54; \
+  __ret_54 = vsetq_lane_p8(vgetq_lane_p8(__s2_54, __p3_54), __s0_54, __p1_54); \
   __ret_54; \
 })
 #else
-#define vcopy_lane_s16(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
-  int16x4_t __s0_55 = __p0_55; \
-  int16x4_t __s2_55 = __p2_55; \
-  int16x4_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 3, 2, 1, 0); \
-  int16x4_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 3, 2, 1, 0); \
-  int16x4_t __ret_55; \
-  __ret_55 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_55, __p3_55), __rev0_55, __p1_55); \
-  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 3, 2, 1, 0); \
+#define vcopyq_laneq_p8(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
+  poly8x16_t __s0_55 = __p0_55; \
+  poly8x16_t __s2_55 = __p2_55; \
+  poly8x16_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __ret_55; \
+  __ret_55 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_55, __p3_55), __rev0_55, __p1_55); \
+  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_55; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p8(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
-  poly8x16_t __s0_56 = __p0_56; \
-  poly8x16_t __s2_56 = __p2_56; \
-  poly8x16_t __ret_56; \
-  __ret_56 = vsetq_lane_p8(vgetq_lane_p8(__s2_56, __p3_56), __s0_56, __p1_56); \
+#define vcopyq_laneq_p16(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
+  poly16x8_t __s0_56 = __p0_56; \
+  poly16x8_t __s2_56 = __p2_56; \
+  poly16x8_t __ret_56; \
+  __ret_56 = vsetq_lane_p16(vgetq_lane_p16(__s2_56, __p3_56), __s0_56, __p1_56); \
   __ret_56; \
 })
 #else
-#define vcopyq_laneq_p8(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
-  poly8x16_t __s0_57 = __p0_57; \
-  poly8x16_t __s2_57 = __p2_57; \
-  poly8x16_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret_57; \
-  __ret_57 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_57, __p3_57), __rev0_57, __p1_57); \
-  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_p16(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
+  poly16x8_t __s0_57 = __p0_57; \
+  poly16x8_t __s2_57 = __p2_57; \
+  poly16x8_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x8_t __ret_57; \
+  __ret_57 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_57, __p3_57), __rev0_57, __p1_57); \
+  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_57; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p16(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
-  poly16x8_t __s0_58 = __p0_58; \
-  poly16x8_t __s2_58 = __p2_58; \
-  poly16x8_t __ret_58; \
-  __ret_58 = vsetq_lane_p16(vgetq_lane_p16(__s2_58, __p3_58), __s0_58, __p1_58); \
+#define vcopyq_laneq_u8(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
+  uint8x16_t __s0_58 = __p0_58; \
+  uint8x16_t __s2_58 = __p2_58; \
+  uint8x16_t __ret_58; \
+  __ret_58 = vsetq_lane_u8(vgetq_lane_u8(__s2_58, __p3_58), __s0_58, __p1_58); \
   __ret_58; \
 })
 #else
-#define vcopyq_laneq_p16(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
-  poly16x8_t __s0_59 = __p0_59; \
-  poly16x8_t __s2_59 = __p2_59; \
-  poly16x8_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret_59; \
-  __ret_59 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_59, __p3_59), __rev0_59, __p1_59); \
-  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_u8(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
+  uint8x16_t __s0_59 = __p0_59; \
+  uint8x16_t __s2_59 = __p2_59; \
+  uint8x16_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_59; \
+  __ret_59 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_59, __p3_59), __rev0_59, __p1_59); \
+  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_59; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u8(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
-  uint8x16_t __s0_60 = __p0_60; \
-  uint8x16_t __s2_60 = __p2_60; \
-  uint8x16_t __ret_60; \
-  __ret_60 = vsetq_lane_u8(vgetq_lane_u8(__s2_60, __p3_60), __s0_60, __p1_60); \
+#define vcopyq_laneq_u32(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
+  uint32x4_t __s0_60 = __p0_60; \
+  uint32x4_t __s2_60 = __p2_60; \
+  uint32x4_t __ret_60; \
+  __ret_60 = vsetq_lane_u32(vgetq_lane_u32(__s2_60, __p3_60), __s0_60, __p1_60); \
   __ret_60; \
 })
 #else
-#define vcopyq_laneq_u8(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
-  uint8x16_t __s0_61 = __p0_61; \
-  uint8x16_t __s2_61 = __p2_61; \
-  uint8x16_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_61; \
-  __ret_61 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_61, __p3_61), __rev0_61, __p1_61); \
-  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_u32(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
+  uint32x4_t __s0_61 = __p0_61; \
+  uint32x4_t __s2_61 = __p2_61; \
+  uint32x4_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 3, 2, 1, 0); \
+  uint32x4_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 3, 2, 1, 0); \
+  uint32x4_t __ret_61; \
+  __ret_61 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_61, __p3_61), __rev0_61, __p1_61); \
+  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 3, 2, 1, 0); \
   __ret_61; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u32(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
-  uint32x4_t __s0_62 = __p0_62; \
-  uint32x4_t __s2_62 = __p2_62; \
-  uint32x4_t __ret_62; \
-  __ret_62 = vsetq_lane_u32(vgetq_lane_u32(__s2_62, __p3_62), __s0_62, __p1_62); \
+#define vcopyq_laneq_u64(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
+  uint64x2_t __s0_62 = __p0_62; \
+  uint64x2_t __s2_62 = __p2_62; \
+  uint64x2_t __ret_62; \
+  __ret_62 = vsetq_lane_u64(vgetq_lane_u64(__s2_62, __p3_62), __s0_62, __p1_62); \
   __ret_62; \
 })
 #else
-#define vcopyq_laneq_u32(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
-  uint32x4_t __s0_63 = __p0_63; \
-  uint32x4_t __s2_63 = __p2_63; \
-  uint32x4_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 3, 2, 1, 0); \
-  uint32x4_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 3, 2, 1, 0); \
-  uint32x4_t __ret_63; \
-  __ret_63 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_63, __p3_63), __rev0_63, __p1_63); \
-  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 3, 2, 1, 0); \
+#define vcopyq_laneq_u64(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
+  uint64x2_t __s0_63 = __p0_63; \
+  uint64x2_t __s2_63 = __p2_63; \
+  uint64x2_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 1, 0); \
+  uint64x2_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 1, 0); \
+  uint64x2_t __ret_63; \
+  __ret_63 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_63, __p3_63), __rev0_63, __p1_63); \
+  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 1, 0); \
   __ret_63; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u64(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
-  uint64x2_t __s0_64 = __p0_64; \
-  uint64x2_t __s2_64 = __p2_64; \
-  uint64x2_t __ret_64; \
-  __ret_64 = vsetq_lane_u64(vgetq_lane_u64(__s2_64, __p3_64), __s0_64, __p1_64); \
+#define vcopyq_laneq_u16(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
+  uint16x8_t __s0_64 = __p0_64; \
+  uint16x8_t __s2_64 = __p2_64; \
+  uint16x8_t __ret_64; \
+  __ret_64 = vsetq_lane_u16(vgetq_lane_u16(__s2_64, __p3_64), __s0_64, __p1_64); \
   __ret_64; \
 })
 #else
-#define vcopyq_laneq_u64(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
-  uint64x2_t __s0_65 = __p0_65; \
-  uint64x2_t __s2_65 = __p2_65; \
-  uint64x2_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 1, 0); \
-  uint64x2_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 1, 0); \
-  uint64x2_t __ret_65; \
-  __ret_65 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_65, __p3_65), __rev0_65, __p1_65); \
-  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 1, 0); \
+#define vcopyq_laneq_u16(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
+  uint16x8_t __s0_65 = __p0_65; \
+  uint16x8_t __s2_65 = __p2_65; \
+  uint16x8_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_65; \
+  __ret_65 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_65, __p3_65), __rev0_65, __p1_65); \
+  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_65; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u16(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
-  uint16x8_t __s0_66 = __p0_66; \
-  uint16x8_t __s2_66 = __p2_66; \
-  uint16x8_t __ret_66; \
-  __ret_66 = vsetq_lane_u16(vgetq_lane_u16(__s2_66, __p3_66), __s0_66, __p1_66); \
+#define vcopyq_laneq_s8(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
+  int8x16_t __s0_66 = __p0_66; \
+  int8x16_t __s2_66 = __p2_66; \
+  int8x16_t __ret_66; \
+  __ret_66 = vsetq_lane_s8(vgetq_lane_s8(__s2_66, __p3_66), __s0_66, __p1_66); \
   __ret_66; \
 })
 #else
-#define vcopyq_laneq_u16(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
-  uint16x8_t __s0_67 = __p0_67; \
-  uint16x8_t __s2_67 = __p2_67; \
-  uint16x8_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_67; \
-  __ret_67 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_67, __p3_67), __rev0_67, __p1_67); \
-  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_s8(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
+  int8x16_t __s0_67 = __p0_67; \
+  int8x16_t __s2_67 = __p2_67; \
+  int8x16_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_67; \
+  __ret_67 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_67, __p3_67), __rev0_67, __p1_67); \
+  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_67; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s8(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
-  int8x16_t __s0_68 = __p0_68; \
-  int8x16_t __s2_68 = __p2_68; \
-  int8x16_t __ret_68; \
-  __ret_68 = vsetq_lane_s8(vgetq_lane_s8(__s2_68, __p3_68), __s0_68, __p1_68); \
+#define vcopyq_laneq_f32(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
+  float32x4_t __s0_68 = __p0_68; \
+  float32x4_t __s2_68 = __p2_68; \
+  float32x4_t __ret_68; \
+  __ret_68 = vsetq_lane_f32(vgetq_lane_f32(__s2_68, __p3_68), __s0_68, __p1_68); \
   __ret_68; \
 })
 #else
-#define vcopyq_laneq_s8(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
-  int8x16_t __s0_69 = __p0_69; \
-  int8x16_t __s2_69 = __p2_69; \
-  int8x16_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_69; \
-  __ret_69 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_69, __p3_69), __rev0_69, __p1_69); \
-  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopyq_laneq_f32(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
+  float32x4_t __s0_69 = __p0_69; \
+  float32x4_t __s2_69 = __p2_69; \
+  float32x4_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 3, 2, 1, 0); \
+  float32x4_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 3, 2, 1, 0); \
+  float32x4_t __ret_69; \
+  __ret_69 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_69, __p3_69), __rev0_69, __p1_69); \
+  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 3, 2, 1, 0); \
   __ret_69; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
-  float32x4_t __s0_70 = __p0_70; \
-  float32x4_t __s2_70 = __p2_70; \
-  float32x4_t __ret_70; \
-  __ret_70 = vsetq_lane_f32(vgetq_lane_f32(__s2_70, __p3_70), __s0_70, __p1_70); \
+#define vcopyq_laneq_s32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
+  int32x4_t __s0_70 = __p0_70; \
+  int32x4_t __s2_70 = __p2_70; \
+  int32x4_t __ret_70; \
+  __ret_70 = vsetq_lane_s32(vgetq_lane_s32(__s2_70, __p3_70), __s0_70, __p1_70); \
   __ret_70; \
 })
 #else
-#define vcopyq_laneq_f32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
-  float32x4_t __s0_71 = __p0_71; \
-  float32x4_t __s2_71 = __p2_71; \
-  float32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
-  float32x4_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 3, 2, 1, 0); \
-  float32x4_t __ret_71; \
-  __ret_71 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_71, __p3_71), __rev0_71, __p1_71); \
+#define vcopyq_laneq_s32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
+  int32x4_t __s0_71 = __p0_71; \
+  int32x4_t __s2_71 = __p2_71; \
+  int32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
+  int32x4_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 3, 2, 1, 0); \
+  int32x4_t __ret_71; \
+  __ret_71 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_71, __p3_71), __rev0_71, __p1_71); \
   __ret_71 = __builtin_shufflevector(__ret_71, __ret_71, 3, 2, 1, 0); \
   __ret_71; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s32(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
-  int32x4_t __s0_72 = __p0_72; \
-  int32x4_t __s2_72 = __p2_72; \
-  int32x4_t __ret_72; \
-  __ret_72 = vsetq_lane_s32(vgetq_lane_s32(__s2_72, __p3_72), __s0_72, __p1_72); \
+#define vcopyq_laneq_s64(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
+  int64x2_t __s0_72 = __p0_72; \
+  int64x2_t __s2_72 = __p2_72; \
+  int64x2_t __ret_72; \
+  __ret_72 = vsetq_lane_s64(vgetq_lane_s64(__s2_72, __p3_72), __s0_72, __p1_72); \
   __ret_72; \
 })
 #else
-#define vcopyq_laneq_s32(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
-  int32x4_t __s0_73 = __p0_73; \
-  int32x4_t __s2_73 = __p2_73; \
-  int32x4_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 3, 2, 1, 0); \
-  int32x4_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 3, 2, 1, 0); \
-  int32x4_t __ret_73; \
-  __ret_73 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_73, __p3_73), __rev0_73, __p1_73); \
-  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 3, 2, 1, 0); \
+#define vcopyq_laneq_s64(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
+  int64x2_t __s0_73 = __p0_73; \
+  int64x2_t __s2_73 = __p2_73; \
+  int64x2_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 1, 0); \
+  int64x2_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 1, 0); \
+  int64x2_t __ret_73; \
+  __ret_73 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_73, __p3_73), __rev0_73, __p1_73); \
+  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 1, 0); \
   __ret_73; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s64(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
-  int64x2_t __s0_74 = __p0_74; \
-  int64x2_t __s2_74 = __p2_74; \
-  int64x2_t __ret_74; \
-  __ret_74 = vsetq_lane_s64(vgetq_lane_s64(__s2_74, __p3_74), __s0_74, __p1_74); \
+#define vcopyq_laneq_s16(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
+  int16x8_t __s0_74 = __p0_74; \
+  int16x8_t __s2_74 = __p2_74; \
+  int16x8_t __ret_74; \
+  __ret_74 = vsetq_lane_s16(vgetq_lane_s16(__s2_74, __p3_74), __s0_74, __p1_74); \
   __ret_74; \
 })
 #else
-#define vcopyq_laneq_s64(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
-  int64x2_t __s0_75 = __p0_75; \
-  int64x2_t __s2_75 = __p2_75; \
-  int64x2_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 1, 0); \
-  int64x2_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 1, 0); \
-  int64x2_t __ret_75; \
-  __ret_75 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_75, __p3_75), __rev0_75, __p1_75); \
-  __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 1, 0); \
+#define vcopyq_laneq_s16(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
+  int16x8_t __s0_75 = __p0_75; \
+  int16x8_t __s2_75 = __p2_75; \
+  int16x8_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_75; \
+  __ret_75 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_75, __p3_75), __rev0_75, __p1_75); \
+  __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_75; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s16(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
-  int16x8_t __s0_76 = __p0_76; \
-  int16x8_t __s2_76 = __p2_76; \
-  int16x8_t __ret_76; \
-  __ret_76 = vsetq_lane_s16(vgetq_lane_s16(__s2_76, __p3_76), __s0_76, __p1_76); \
+#define vcopy_laneq_p8(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
+  poly8x8_t __s0_76 = __p0_76; \
+  poly8x16_t __s2_76 = __p2_76; \
+  poly8x8_t __ret_76; \
+  __ret_76 = vset_lane_p8(vgetq_lane_p8(__s2_76, __p3_76), __s0_76, __p1_76); \
   __ret_76; \
 })
 #else
-#define vcopyq_laneq_s16(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
-  int16x8_t __s0_77 = __p0_77; \
-  int16x8_t __s2_77 = __p2_77; \
-  int16x8_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_77; \
-  __ret_77 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_77, __p3_77), __rev0_77, __p1_77); \
+#define vcopy_laneq_p8(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
+  poly8x8_t __s0_77 = __p0_77; \
+  poly8x16_t __s2_77 = __p2_77; \
+  poly8x8_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x16_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly8x8_t __ret_77; \
+  __ret_77 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_77, __p3_77), __rev0_77, __p1_77); \
   __ret_77 = __builtin_shufflevector(__ret_77, __ret_77, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_77; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p8(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
-  poly8x8_t __s0_78 = __p0_78; \
-  poly8x16_t __s2_78 = __p2_78; \
-  poly8x8_t __ret_78; \
-  __ret_78 = vset_lane_p8(vgetq_lane_p8(__s2_78, __p3_78), __s0_78, __p1_78); \
+#define vcopy_laneq_p16(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
+  poly16x4_t __s0_78 = __p0_78; \
+  poly16x8_t __s2_78 = __p2_78; \
+  poly16x4_t __ret_78; \
+  __ret_78 = vset_lane_p16(vgetq_lane_p16(__s2_78, __p3_78), __s0_78, __p1_78); \
   __ret_78; \
 })
 #else
-#define vcopy_laneq_p8(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
-  poly8x8_t __s0_79 = __p0_79; \
-  poly8x16_t __s2_79 = __p2_79; \
-  poly8x8_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret_79; \
-  __ret_79 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_79, __p3_79), __rev0_79, __p1_79); \
-  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_p16(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
+  poly16x4_t __s0_79 = __p0_79; \
+  poly16x8_t __s2_79 = __p2_79; \
+  poly16x4_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 3, 2, 1, 0); \
+  poly16x8_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 7, 6, 5, 4, 3, 2, 1, 0); \
+  poly16x4_t __ret_79; \
+  __ret_79 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_79, __p3_79), __rev0_79, __p1_79); \
+  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 3, 2, 1, 0); \
   __ret_79; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p16(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
-  poly16x4_t __s0_80 = __p0_80; \
-  poly16x8_t __s2_80 = __p2_80; \
-  poly16x4_t __ret_80; \
-  __ret_80 = vset_lane_p16(vgetq_lane_p16(__s2_80, __p3_80), __s0_80, __p1_80); \
+#define vcopy_laneq_u8(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
+  uint8x8_t __s0_80 = __p0_80; \
+  uint8x16_t __s2_80 = __p2_80; \
+  uint8x8_t __ret_80; \
+  __ret_80 = vset_lane_u8(vgetq_lane_u8(__s2_80, __p3_80), __s0_80, __p1_80); \
   __ret_80; \
 })
 #else
-#define vcopy_laneq_p16(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
-  poly16x4_t __s0_81 = __p0_81; \
-  poly16x8_t __s2_81 = __p2_81; \
-  poly16x4_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 3, 2, 1, 0); \
-  poly16x8_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x4_t __ret_81; \
-  __ret_81 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_81, __p3_81), __rev0_81, __p1_81); \
-  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 3, 2, 1, 0); \
+#define vcopy_laneq_u8(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
+  uint8x8_t __s0_81 = __p0_81; \
+  uint8x16_t __s2_81 = __p2_81; \
+  uint8x8_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x8_t __ret_81; \
+  __ret_81 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_81, __p3_81), __rev0_81, __p1_81); \
+  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_81; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u8(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
-  uint8x8_t __s0_82 = __p0_82; \
-  uint8x16_t __s2_82 = __p2_82; \
-  uint8x8_t __ret_82; \
-  __ret_82 = vset_lane_u8(vgetq_lane_u8(__s2_82, __p3_82), __s0_82, __p1_82); \
+#define vcopy_laneq_u32(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
+  uint32x2_t __s0_82 = __p0_82; \
+  uint32x4_t __s2_82 = __p2_82; \
+  uint32x2_t __ret_82; \
+  __ret_82 = vset_lane_u32(vgetq_lane_u32(__s2_82, __p3_82), __s0_82, __p1_82); \
   __ret_82; \
 })
 #else
-#define vcopy_laneq_u8(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
-  uint8x8_t __s0_83 = __p0_83; \
-  uint8x16_t __s2_83 = __p2_83; \
-  uint8x8_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret_83; \
-  __ret_83 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_83, __p3_83), __rev0_83, __p1_83); \
-  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_u32(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
+  uint32x2_t __s0_83 = __p0_83; \
+  uint32x4_t __s2_83 = __p2_83; \
+  uint32x2_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 1, 0); \
+  uint32x4_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 3, 2, 1, 0); \
+  uint32x2_t __ret_83; \
+  __ret_83 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_83, __p3_83), __rev0_83, __p1_83); \
+  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 1, 0); \
   __ret_83; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u32(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
-  uint32x2_t __s0_84 = __p0_84; \
-  uint32x4_t __s2_84 = __p2_84; \
-  uint32x2_t __ret_84; \
-  __ret_84 = vset_lane_u32(vgetq_lane_u32(__s2_84, __p3_84), __s0_84, __p1_84); \
+#define vcopy_laneq_u64(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
+  uint64x1_t __s0_84 = __p0_84; \
+  uint64x2_t __s2_84 = __p2_84; \
+  uint64x1_t __ret_84; \
+  __ret_84 = vset_lane_u64(vgetq_lane_u64(__s2_84, __p3_84), __s0_84, __p1_84); \
   __ret_84; \
 })
 #else
-#define vcopy_laneq_u32(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
-  uint32x2_t __s0_85 = __p0_85; \
-  uint32x4_t __s2_85 = __p2_85; \
-  uint32x2_t __rev0_85;  __rev0_85 = __builtin_shufflevector(__s0_85, __s0_85, 1, 0); \
-  uint32x4_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 3, 2, 1, 0); \
-  uint32x2_t __ret_85; \
-  __ret_85 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_85, __p3_85), __rev0_85, __p1_85); \
-  __ret_85 = __builtin_shufflevector(__ret_85, __ret_85, 1, 0); \
+#define vcopy_laneq_u64(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
+  uint64x1_t __s0_85 = __p0_85; \
+  uint64x2_t __s2_85 = __p2_85; \
+  uint64x2_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 1, 0); \
+  uint64x1_t __ret_85; \
+  __ret_85 = vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_85, __p3_85), __s0_85, __p1_85); \
   __ret_85; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u64(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
-  uint64x1_t __s0_86 = __p0_86; \
-  uint64x2_t __s2_86 = __p2_86; \
-  uint64x1_t __ret_86; \
-  __ret_86 = vset_lane_u64(vgetq_lane_u64(__s2_86, __p3_86), __s0_86, __p1_86); \
+#define vcopy_laneq_u16(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
+  uint16x4_t __s0_86 = __p0_86; \
+  uint16x8_t __s2_86 = __p2_86; \
+  uint16x4_t __ret_86; \
+  __ret_86 = vset_lane_u16(vgetq_lane_u16(__s2_86, __p3_86), __s0_86, __p1_86); \
   __ret_86; \
 })
 #else
-#define vcopy_laneq_u64(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
-  uint64x1_t __s0_87 = __p0_87; \
-  uint64x2_t __s2_87 = __p2_87; \
-  uint64x2_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 1, 0); \
-  uint64x1_t __ret_87; \
-  __ret_87 = __noswap_vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_87, __p3_87), __s0_87, __p1_87); \
+#define vcopy_laneq_u16(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
+  uint16x4_t __s0_87 = __p0_87; \
+  uint16x8_t __s2_87 = __p2_87; \
+  uint16x4_t __rev0_87;  __rev0_87 = __builtin_shufflevector(__s0_87, __s0_87, 3, 2, 1, 0); \
+  uint16x8_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x4_t __ret_87; \
+  __ret_87 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_87, __p3_87), __rev0_87, __p1_87); \
+  __ret_87 = __builtin_shufflevector(__ret_87, __ret_87, 3, 2, 1, 0); \
   __ret_87; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u16(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
-  uint16x4_t __s0_88 = __p0_88; \
-  uint16x8_t __s2_88 = __p2_88; \
-  uint16x4_t __ret_88; \
-  __ret_88 = vset_lane_u16(vgetq_lane_u16(__s2_88, __p3_88), __s0_88, __p1_88); \
+#define vcopy_laneq_s8(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
+  int8x8_t __s0_88 = __p0_88; \
+  int8x16_t __s2_88 = __p2_88; \
+  int8x8_t __ret_88; \
+  __ret_88 = vset_lane_s8(vgetq_lane_s8(__s2_88, __p3_88), __s0_88, __p1_88); \
   __ret_88; \
 })
 #else
-#define vcopy_laneq_u16(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
-  uint16x4_t __s0_89 = __p0_89; \
-  uint16x8_t __s2_89 = __p2_89; \
-  uint16x4_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 3, 2, 1, 0); \
-  uint16x8_t __rev2_89;  __rev2_89 = __builtin_shufflevector(__s2_89, __s2_89, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret_89; \
-  __ret_89 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_89, __p3_89), __rev0_89, __p1_89); \
-  __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 3, 2, 1, 0); \
+#define vcopy_laneq_s8(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
+  int8x8_t __s0_89 = __p0_89; \
+  int8x16_t __s2_89 = __p2_89; \
+  int8x8_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __rev2_89;  __rev2_89 = __builtin_shufflevector(__s2_89, __s2_89, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x8_t __ret_89; \
+  __ret_89 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_89, __p3_89), __rev0_89, __p1_89); \
+  __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_89; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s8(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
-  int8x8_t __s0_90 = __p0_90; \
-  int8x16_t __s2_90 = __p2_90; \
-  int8x8_t __ret_90; \
-  __ret_90 = vset_lane_s8(vgetq_lane_s8(__s2_90, __p3_90), __s0_90, __p1_90); \
+#define vcopy_laneq_f32(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
+  float32x2_t __s0_90 = __p0_90; \
+  float32x4_t __s2_90 = __p2_90; \
+  float32x2_t __ret_90; \
+  __ret_90 = vset_lane_f32(vgetq_lane_f32(__s2_90, __p3_90), __s0_90, __p1_90); \
   __ret_90; \
 })
 #else
-#define vcopy_laneq_s8(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
-  int8x8_t __s0_91 = __p0_91; \
-  int8x16_t __s2_91 = __p2_91; \
-  int8x8_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_91; \
-  __ret_91 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_91, __p3_91), __rev0_91, __p1_91); \
-  __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 7, 6, 5, 4, 3, 2, 1, 0); \
+#define vcopy_laneq_f32(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
+  float32x2_t __s0_91 = __p0_91; \
+  float32x4_t __s2_91 = __p2_91; \
+  float32x2_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 1, 0); \
+  float32x4_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 3, 2, 1, 0); \
+  float32x2_t __ret_91; \
+  __ret_91 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_91, __p3_91), __rev0_91, __p1_91); \
+  __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 1, 0); \
   __ret_91; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f32(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
-  float32x2_t __s0_92 = __p0_92; \
-  float32x4_t __s2_92 = __p2_92; \
-  float32x2_t __ret_92; \
-  __ret_92 = vset_lane_f32(vgetq_lane_f32(__s2_92, __p3_92), __s0_92, __p1_92); \
+#define vcopy_laneq_s32(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
+  int32x2_t __s0_92 = __p0_92; \
+  int32x4_t __s2_92 = __p2_92; \
+  int32x2_t __ret_92; \
+  __ret_92 = vset_lane_s32(vgetq_lane_s32(__s2_92, __p3_92), __s0_92, __p1_92); \
   __ret_92; \
 })
 #else
-#define vcopy_laneq_f32(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
-  float32x2_t __s0_93 = __p0_93; \
-  float32x4_t __s2_93 = __p2_93; \
-  float32x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
-  float32x4_t __rev2_93;  __rev2_93 = __builtin_shufflevector(__s2_93, __s2_93, 3, 2, 1, 0); \
-  float32x2_t __ret_93; \
-  __ret_93 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_93, __p3_93), __rev0_93, __p1_93); \
+#define vcopy_laneq_s32(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
+  int32x2_t __s0_93 = __p0_93; \
+  int32x4_t __s2_93 = __p2_93; \
+  int32x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
+  int32x4_t __rev2_93;  __rev2_93 = __builtin_shufflevector(__s2_93, __s2_93, 3, 2, 1, 0); \
+  int32x2_t __ret_93; \
+  __ret_93 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_93, __p3_93), __rev0_93, __p1_93); \
   __ret_93 = __builtin_shufflevector(__ret_93, __ret_93, 1, 0); \
   __ret_93; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s32(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
-  int32x2_t __s0_94 = __p0_94; \
-  int32x4_t __s2_94 = __p2_94; \
-  int32x2_t __ret_94; \
-  __ret_94 = vset_lane_s32(vgetq_lane_s32(__s2_94, __p3_94), __s0_94, __p1_94); \
+#define vcopy_laneq_s64(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
+  int64x1_t __s0_94 = __p0_94; \
+  int64x2_t __s2_94 = __p2_94; \
+  int64x1_t __ret_94; \
+  __ret_94 = vset_lane_s64(vgetq_lane_s64(__s2_94, __p3_94), __s0_94, __p1_94); \
   __ret_94; \
 })
 #else
-#define vcopy_laneq_s32(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
-  int32x2_t __s0_95 = __p0_95; \
-  int32x4_t __s2_95 = __p2_95; \
-  int32x2_t __rev0_95;  __rev0_95 = __builtin_shufflevector(__s0_95, __s0_95, 1, 0); \
-  int32x4_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 3, 2, 1, 0); \
-  int32x2_t __ret_95; \
-  __ret_95 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_95, __p3_95), __rev0_95, __p1_95); \
-  __ret_95 = __builtin_shufflevector(__ret_95, __ret_95, 1, 0); \
+#define vcopy_laneq_s64(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
+  int64x1_t __s0_95 = __p0_95; \
+  int64x2_t __s2_95 = __p2_95; \
+  int64x2_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 1, 0); \
+  int64x1_t __ret_95; \
+  __ret_95 = vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_95, __p3_95), __s0_95, __p1_95); \
   __ret_95; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s64(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
-  int64x1_t __s0_96 = __p0_96; \
-  int64x2_t __s2_96 = __p2_96; \
-  int64x1_t __ret_96; \
-  __ret_96 = vset_lane_s64(vgetq_lane_s64(__s2_96, __p3_96), __s0_96, __p1_96); \
+#define vcopy_laneq_s16(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
+  int16x4_t __s0_96 = __p0_96; \
+  int16x8_t __s2_96 = __p2_96; \
+  int16x4_t __ret_96; \
+  __ret_96 = vset_lane_s16(vgetq_lane_s16(__s2_96, __p3_96), __s0_96, __p1_96); \
   __ret_96; \
 })
 #else
-#define vcopy_laneq_s64(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
-  int64x1_t __s0_97 = __p0_97; \
-  int64x2_t __s2_97 = __p2_97; \
-  int64x2_t __rev2_97;  __rev2_97 = __builtin_shufflevector(__s2_97, __s2_97, 1, 0); \
-  int64x1_t __ret_97; \
-  __ret_97 = __noswap_vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_97, __p3_97), __s0_97, __p1_97); \
+#define vcopy_laneq_s16(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
+  int16x4_t __s0_97 = __p0_97; \
+  int16x8_t __s2_97 = __p2_97; \
+  int16x4_t __rev0_97;  __rev0_97 = __builtin_shufflevector(__s0_97, __s0_97, 3, 2, 1, 0); \
+  int16x8_t __rev2_97;  __rev2_97 = __builtin_shufflevector(__s2_97, __s2_97, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x4_t __ret_97; \
+  __ret_97 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_97, __p3_97), __rev0_97, __p1_97); \
+  __ret_97 = __builtin_shufflevector(__ret_97, __ret_97, 3, 2, 1, 0); \
   __ret_97; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s16(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
-  int16x4_t __s0_98 = __p0_98; \
-  int16x8_t __s2_98 = __p2_98; \
-  int16x4_t __ret_98; \
-  __ret_98 = vset_lane_s16(vgetq_lane_s16(__s2_98, __p3_98), __s0_98, __p1_98); \
-  __ret_98; \
-})
-#else
-#define vcopy_laneq_s16(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
-  int16x4_t __s0_99 = __p0_99; \
-  int16x8_t __s2_99 = __p2_99; \
-  int16x4_t __rev0_99;  __rev0_99 = __builtin_shufflevector(__s0_99, __s0_99, 3, 2, 1, 0); \
-  int16x8_t __rev2_99;  __rev2_99 = __builtin_shufflevector(__s2_99, __s2_99, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret_99; \
-  __ret_99 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_99, __p3_99), __rev0_99, __p1_99); \
-  __ret_99 = __builtin_shufflevector(__ret_99, __ret_99, 3, 2, 1, 0); \
-  __ret_99; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vcreate_p64(uint64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai poly64x1_t vcreate_p64(uint64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcreate_f64(uint64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t)(__p0);
   return __ret;
 }
-#else
-__ai float64x1_t vcreate_f64(uint64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvts_f32_s32(int32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvts_f32_s32(int32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvts_f32_u32(uint32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvts_f32_u32(uint32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
   float32x2_t __ret;
@@ -52270,34 +44324,16 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vcvtd_f64_s64(int64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vcvtd_f64_s64(int64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vcvtd_f64_u64(uint64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
   float64x2_t __ret;
@@ -52330,34 +44366,16 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
   float64x2_t __ret;
@@ -52445,38 +44463,18 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
   uint32_t __s0 = __p0; \
   float32_t __ret; \
   __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   float32_t __ret; \
   __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64x2_t __s0 = __p0; \
@@ -52513,86 +44511,36 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
   float32_t __s0 = __p0; \
   int32_t __ret; \
   __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -52611,54 +44559,24 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   int64x1_t __ret; \
   __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
   __ret; \
 })
-#else
-#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
   float64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
   float32_t __s0 = __p0; \
   uint32_t __ret; \
   __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -52677,66 +44595,28 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   uint64x1_t __ret; \
   __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
   __ret; \
 })
-#else
-#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
   float64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvts_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvts_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
   int64x2_t __ret;
@@ -52753,48 +44633,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvts_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvts_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
   uint64x2_t __ret;
@@ -52811,258 +44664,96 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtas_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtas_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtad_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtad_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtas_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtad_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtms_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtms_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtmd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtms_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtns_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtns_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtnd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtns_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vcvtps_s32_f32(float32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vcvtps_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vcvtpd_s64_f64(float64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vcvtps_u32_f32(float32_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
   return __ret;
 }
-#else
-__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vcvtxd_f32_f64(float64_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
   return __ret;
 }
-#else
-__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
   float32x2_t __ret;
@@ -53135,20 +44826,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 / __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
   float32x2_t __ret;
@@ -53234,22 +44916,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
   uint64x1_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vduph_lane_u16(__p0, __p1) __extension__ ({ \
   uint16x4_t __s0 = __p0; \
@@ -53284,22 +44956,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdups_lane_f32(__p0, __p1) __extension__ ({ \
   float32x2_t __s0 = __p0; \
@@ -53334,22 +44996,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
   int64x1_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vduph_lane_s16(__p0, __p1) __extension__ ({ \
   int16x4_t __s0 = __p0; \
@@ -53367,22 +45019,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
@@ -53435,22 +45077,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vdup_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __ret; \
   __ret = __builtin_shufflevector(__s0, __s0, __p1); \
   __ret; \
 })
-#else
-#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vdup_lane_f16(__p0, __p1) __extension__ ({ \
   float16x4_t __s0 = __p0; \
@@ -54173,20 +45805,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vdup_n_p64(poly64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
   poly64x2_t __ret;
@@ -54217,21 +45840,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vdup_n_f64(float64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai float64x1_t vdup_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vext_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -54239,16 +45852,6 @@
   __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -54291,7 +45894,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vext_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -54299,16 +45901,6 @@
   __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
   float64x2_t __ret;
@@ -54332,26 +45924,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-__ai float64x1_t __noswap_vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
   float64_t __s0 = __p0; \
   float64_t __s1 = __p1; \
@@ -54360,25 +45937,6 @@
   __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
   __ret; \
 })
-#else
-#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
   float32_t __s0 = __p0; \
@@ -54471,7 +46029,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -54480,25 +46037,6 @@
   __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
   __ret; \
 })
-#else
-#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#define __noswap_vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
   float32x2_t __s0 = __p0; \
@@ -54731,20 +46269,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, __p1, (float64x1_t) {__p2});
   return __ret;
 }
-#else
-__ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, __p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
   float64x2_t __ret;
@@ -54763,288 +46292,255 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, -__p1, __p2);
   return __ret;
 }
-#else
-__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, __p2);
-  return __ret;
-}
-#endif
-
+#define vfmsd_lane_f64(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
+  float64_t __s0_98 = __p0_98; \
+  float64_t __s1_98 = __p1_98; \
+  float64x1_t __s2_98 = __p2_98; \
+  float64_t __ret_98; \
+  __ret_98 = vfmad_lane_f64(__s0_98, -__s1_98, __s2_98, __p3_98); \
+  __ret_98; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vfmsd_lane_f64(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
-  float64_t __s0_100 = __p0_100; \
-  float64_t __s1_100 = __p1_100; \
-  float64x1_t __s2_100 = __p2_100; \
-  float64_t __ret_100; \
-  __ret_100 = vfmad_lane_f64(__s0_100, -__s1_100, __s2_100, __p3_100); \
+#define vfmss_lane_f32(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
+  float32_t __s0_99 = __p0_99; \
+  float32_t __s1_99 = __p1_99; \
+  float32x2_t __s2_99 = __p2_99; \
+  float32_t __ret_99; \
+  __ret_99 = vfmas_lane_f32(__s0_99, -__s1_99, __s2_99, __p3_99); \
+  __ret_99; \
+})
+#else
+#define vfmss_lane_f32(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
+  float32_t __s0_100 = __p0_100; \
+  float32_t __s1_100 = __p1_100; \
+  float32x2_t __s2_100 = __p2_100; \
+  float32x2_t __rev2_100;  __rev2_100 = __builtin_shufflevector(__s2_100, __s2_100, 1, 0); \
+  float32_t __ret_100; \
+  __ret_100 = __noswap_vfmas_lane_f32(__s0_100, -__s1_100, __rev2_100, __p3_100); \
   __ret_100; \
 })
-#else
-#define vfmsd_lane_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
-  float64_t __s0_101 = __p0_101; \
-  float64_t __s1_101 = __p1_101; \
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
+  float64x2_t __s0_101 = __p0_101; \
+  float64x2_t __s1_101 = __p1_101; \
   float64x1_t __s2_101 = __p2_101; \
-  float64_t __ret_101; \
-  __ret_101 = __noswap_vfmad_lane_f64(__s0_101, -__s1_101, __s2_101, __p3_101); \
+  float64x2_t __ret_101; \
+  __ret_101 = vfmaq_lane_f64(__s0_101, -__s1_101, __s2_101, __p3_101); \
   __ret_101; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmss_lane_f32(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
-  float32_t __s0_102 = __p0_102; \
-  float32_t __s1_102 = __p1_102; \
-  float32x2_t __s2_102 = __p2_102; \
-  float32_t __ret_102; \
-  __ret_102 = vfmas_lane_f32(__s0_102, -__s1_102, __s2_102, __p3_102); \
+#else
+#define vfmsq_lane_f64(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
+  float64x2_t __s0_102 = __p0_102; \
+  float64x2_t __s1_102 = __p1_102; \
+  float64x1_t __s2_102 = __p2_102; \
+  float64x2_t __rev0_102;  __rev0_102 = __builtin_shufflevector(__s0_102, __s0_102, 1, 0); \
+  float64x2_t __rev1_102;  __rev1_102 = __builtin_shufflevector(__s1_102, __s1_102, 1, 0); \
+  float64x2_t __ret_102; \
+  __ret_102 = __noswap_vfmaq_lane_f64(__rev0_102, -__rev1_102, __s2_102, __p3_102); \
+  __ret_102 = __builtin_shufflevector(__ret_102, __ret_102, 1, 0); \
   __ret_102; \
 })
-#else
-#define vfmss_lane_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
-  float32_t __s0_103 = __p0_103; \
-  float32_t __s1_103 = __p1_103; \
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmsq_lane_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
+  float32x4_t __s0_103 = __p0_103; \
+  float32x4_t __s1_103 = __p1_103; \
   float32x2_t __s2_103 = __p2_103; \
-  float32x2_t __rev2_103;  __rev2_103 = __builtin_shufflevector(__s2_103, __s2_103, 1, 0); \
-  float32_t __ret_103; \
-  __ret_103 = __noswap_vfmas_lane_f32(__s0_103, -__s1_103, __rev2_103, __p3_103); \
+  float32x4_t __ret_103; \
+  __ret_103 = vfmaq_lane_f32(__s0_103, -__s1_103, __s2_103, __p3_103); \
   __ret_103; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f64(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
-  float64x2_t __s0_104 = __p0_104; \
-  float64x2_t __s1_104 = __p1_104; \
-  float64x1_t __s2_104 = __p2_104; \
-  float64x2_t __ret_104; \
-  __ret_104 = vfmaq_lane_f64(__s0_104, -__s1_104, __s2_104, __p3_104); \
+#else
+#define vfmsq_lane_f32(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
+  float32x4_t __s0_104 = __p0_104; \
+  float32x4_t __s1_104 = __p1_104; \
+  float32x2_t __s2_104 = __p2_104; \
+  float32x4_t __rev0_104;  __rev0_104 = __builtin_shufflevector(__s0_104, __s0_104, 3, 2, 1, 0); \
+  float32x4_t __rev1_104;  __rev1_104 = __builtin_shufflevector(__s1_104, __s1_104, 3, 2, 1, 0); \
+  float32x2_t __rev2_104;  __rev2_104 = __builtin_shufflevector(__s2_104, __s2_104, 1, 0); \
+  float32x4_t __ret_104; \
+  __ret_104 = __noswap_vfmaq_lane_f32(__rev0_104, -__rev1_104, __rev2_104, __p3_104); \
+  __ret_104 = __builtin_shufflevector(__ret_104, __ret_104, 3, 2, 1, 0); \
   __ret_104; \
 })
-#else
-#define vfmsq_lane_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
-  float64x2_t __s0_105 = __p0_105; \
-  float64x2_t __s1_105 = __p1_105; \
-  float64x1_t __s2_105 = __p2_105; \
-  float64x2_t __rev0_105;  __rev0_105 = __builtin_shufflevector(__s0_105, __s0_105, 1, 0); \
-  float64x2_t __rev1_105;  __rev1_105 = __builtin_shufflevector(__s1_105, __s1_105, 1, 0); \
-  float64x2_t __ret_105; \
-  __ret_105 = __noswap_vfmaq_lane_f64(__rev0_105, -__rev1_105, __s2_105, __p3_105); \
-  __ret_105 = __builtin_shufflevector(__ret_105, __ret_105, 1, 0); \
-  __ret_105; \
-})
 #endif
 
+#define vfms_lane_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
+  float64x1_t __s0_105 = __p0_105; \
+  float64x1_t __s1_105 = __p1_105; \
+  float64x1_t __s2_105 = __p2_105; \
+  float64x1_t __ret_105; \
+  __ret_105 = vfma_lane_f64(__s0_105, -__s1_105, __s2_105, __p3_105); \
+  __ret_105; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
-  float32x4_t __s0_106 = __p0_106; \
-  float32x4_t __s1_106 = __p1_106; \
+#define vfms_lane_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
+  float32x2_t __s0_106 = __p0_106; \
+  float32x2_t __s1_106 = __p1_106; \
   float32x2_t __s2_106 = __p2_106; \
-  float32x4_t __ret_106; \
-  __ret_106 = vfmaq_lane_f32(__s0_106, -__s1_106, __s2_106, __p3_106); \
+  float32x2_t __ret_106; \
+  __ret_106 = vfma_lane_f32(__s0_106, -__s1_106, __s2_106, __p3_106); \
   __ret_106; \
 })
 #else
-#define vfmsq_lane_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
-  float32x4_t __s0_107 = __p0_107; \
-  float32x4_t __s1_107 = __p1_107; \
+#define vfms_lane_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
+  float32x2_t __s0_107 = __p0_107; \
+  float32x2_t __s1_107 = __p1_107; \
   float32x2_t __s2_107 = __p2_107; \
-  float32x4_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 3, 2, 1, 0); \
-  float32x4_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 3, 2, 1, 0); \
+  float32x2_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 1, 0); \
+  float32x2_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 1, 0); \
   float32x2_t __rev2_107;  __rev2_107 = __builtin_shufflevector(__s2_107, __s2_107, 1, 0); \
-  float32x4_t __ret_107; \
-  __ret_107 = __noswap_vfmaq_lane_f32(__rev0_107, -__rev1_107, __rev2_107, __p3_107); \
-  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 3, 2, 1, 0); \
+  float32x2_t __ret_107; \
+  __ret_107 = __noswap_vfma_lane_f32(__rev0_107, -__rev1_107, __rev2_107, __p3_107); \
+  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 1, 0); \
   __ret_107; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
-  float64x1_t __s0_108 = __p0_108; \
-  float64x1_t __s1_108 = __p1_108; \
-  float64x1_t __s2_108 = __p2_108; \
-  float64x1_t __ret_108; \
-  __ret_108 = vfma_lane_f64(__s0_108, -__s1_108, __s2_108, __p3_108); \
+#define vfmsd_laneq_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
+  float64_t __s0_108 = __p0_108; \
+  float64_t __s1_108 = __p1_108; \
+  float64x2_t __s2_108 = __p2_108; \
+  float64_t __ret_108; \
+  __ret_108 = vfmad_laneq_f64(__s0_108, -__s1_108, __s2_108, __p3_108); \
   __ret_108; \
 })
 #else
-#define vfms_lane_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
-  float64x1_t __s0_109 = __p0_109; \
-  float64x1_t __s1_109 = __p1_109; \
-  float64x1_t __s2_109 = __p2_109; \
-  float64x1_t __ret_109; \
-  __ret_109 = __noswap_vfma_lane_f64(__s0_109, -__s1_109, __s2_109, __p3_109); \
+#define vfmsd_laneq_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
+  float64_t __s0_109 = __p0_109; \
+  float64_t __s1_109 = __p1_109; \
+  float64x2_t __s2_109 = __p2_109; \
+  float64x2_t __rev2_109;  __rev2_109 = __builtin_shufflevector(__s2_109, __s2_109, 1, 0); \
+  float64_t __ret_109; \
+  __ret_109 = __noswap_vfmad_laneq_f64(__s0_109, -__s1_109, __rev2_109, __p3_109); \
   __ret_109; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
-  float32x2_t __s0_110 = __p0_110; \
-  float32x2_t __s1_110 = __p1_110; \
-  float32x2_t __s2_110 = __p2_110; \
-  float32x2_t __ret_110; \
-  __ret_110 = vfma_lane_f32(__s0_110, -__s1_110, __s2_110, __p3_110); \
+#define vfmss_laneq_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
+  float32_t __s0_110 = __p0_110; \
+  float32_t __s1_110 = __p1_110; \
+  float32x4_t __s2_110 = __p2_110; \
+  float32_t __ret_110; \
+  __ret_110 = vfmas_laneq_f32(__s0_110, -__s1_110, __s2_110, __p3_110); \
   __ret_110; \
 })
 #else
-#define vfms_lane_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
-  float32x2_t __s0_111 = __p0_111; \
-  float32x2_t __s1_111 = __p1_111; \
-  float32x2_t __s2_111 = __p2_111; \
-  float32x2_t __rev0_111;  __rev0_111 = __builtin_shufflevector(__s0_111, __s0_111, 1, 0); \
-  float32x2_t __rev1_111;  __rev1_111 = __builtin_shufflevector(__s1_111, __s1_111, 1, 0); \
-  float32x2_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 1, 0); \
-  float32x2_t __ret_111; \
-  __ret_111 = __noswap_vfma_lane_f32(__rev0_111, -__rev1_111, __rev2_111, __p3_111); \
-  __ret_111 = __builtin_shufflevector(__ret_111, __ret_111, 1, 0); \
+#define vfmss_laneq_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
+  float32_t __s0_111 = __p0_111; \
+  float32_t __s1_111 = __p1_111; \
+  float32x4_t __s2_111 = __p2_111; \
+  float32x4_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 3, 2, 1, 0); \
+  float32_t __ret_111; \
+  __ret_111 = __noswap_vfmas_laneq_f32(__s0_111, -__s1_111, __rev2_111, __p3_111); \
   __ret_111; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsd_laneq_f64(__p0_112, __p1_112, __p2_112, __p3_112) __extension__ ({ \
-  float64_t __s0_112 = __p0_112; \
-  float64_t __s1_112 = __p1_112; \
+#define vfmsq_laneq_f64(__p0_112, __p1_112, __p2_112, __p3_112) __extension__ ({ \
+  float64x2_t __s0_112 = __p0_112; \
+  float64x2_t __s1_112 = __p1_112; \
   float64x2_t __s2_112 = __p2_112; \
-  float64_t __ret_112; \
-  __ret_112 = vfmad_laneq_f64(__s0_112, -__s1_112, __s2_112, __p3_112); \
+  float64x2_t __ret_112; \
+  __ret_112 = vfmaq_laneq_f64(__s0_112, -__s1_112, __s2_112, __p3_112); \
   __ret_112; \
 })
 #else
-#define vfmsd_laneq_f64(__p0_113, __p1_113, __p2_113, __p3_113) __extension__ ({ \
-  float64_t __s0_113 = __p0_113; \
-  float64_t __s1_113 = __p1_113; \
+#define vfmsq_laneq_f64(__p0_113, __p1_113, __p2_113, __p3_113) __extension__ ({ \
+  float64x2_t __s0_113 = __p0_113; \
+  float64x2_t __s1_113 = __p1_113; \
   float64x2_t __s2_113 = __p2_113; \
+  float64x2_t __rev0_113;  __rev0_113 = __builtin_shufflevector(__s0_113, __s0_113, 1, 0); \
+  float64x2_t __rev1_113;  __rev1_113 = __builtin_shufflevector(__s1_113, __s1_113, 1, 0); \
   float64x2_t __rev2_113;  __rev2_113 = __builtin_shufflevector(__s2_113, __s2_113, 1, 0); \
-  float64_t __ret_113; \
-  __ret_113 = __noswap_vfmad_laneq_f64(__s0_113, -__s1_113, __rev2_113, __p3_113); \
+  float64x2_t __ret_113; \
+  __ret_113 = __noswap_vfmaq_laneq_f64(__rev0_113, -__rev1_113, __rev2_113, __p3_113); \
+  __ret_113 = __builtin_shufflevector(__ret_113, __ret_113, 1, 0); \
   __ret_113; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmss_laneq_f32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
-  float32_t __s0_114 = __p0_114; \
-  float32_t __s1_114 = __p1_114; \
+#define vfmsq_laneq_f32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
+  float32x4_t __s0_114 = __p0_114; \
+  float32x4_t __s1_114 = __p1_114; \
   float32x4_t __s2_114 = __p2_114; \
-  float32_t __ret_114; \
-  __ret_114 = vfmas_laneq_f32(__s0_114, -__s1_114, __s2_114, __p3_114); \
+  float32x4_t __ret_114; \
+  __ret_114 = vfmaq_laneq_f32(__s0_114, -__s1_114, __s2_114, __p3_114); \
   __ret_114; \
 })
 #else
-#define vfmss_laneq_f32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
-  float32_t __s0_115 = __p0_115; \
-  float32_t __s1_115 = __p1_115; \
+#define vfmsq_laneq_f32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
+  float32x4_t __s0_115 = __p0_115; \
+  float32x4_t __s1_115 = __p1_115; \
   float32x4_t __s2_115 = __p2_115; \
+  float32x4_t __rev0_115;  __rev0_115 = __builtin_shufflevector(__s0_115, __s0_115, 3, 2, 1, 0); \
+  float32x4_t __rev1_115;  __rev1_115 = __builtin_shufflevector(__s1_115, __s1_115, 3, 2, 1, 0); \
   float32x4_t __rev2_115;  __rev2_115 = __builtin_shufflevector(__s2_115, __s2_115, 3, 2, 1, 0); \
-  float32_t __ret_115; \
-  __ret_115 = __noswap_vfmas_laneq_f32(__s0_115, -__s1_115, __rev2_115, __p3_115); \
+  float32x4_t __ret_115; \
+  __ret_115 = __noswap_vfmaq_laneq_f32(__rev0_115, -__rev1_115, __rev2_115, __p3_115); \
+  __ret_115 = __builtin_shufflevector(__ret_115, __ret_115, 3, 2, 1, 0); \
   __ret_115; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f64(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
-  float64x2_t __s0_116 = __p0_116; \
-  float64x2_t __s1_116 = __p1_116; \
+#define vfms_laneq_f64(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
+  float64x1_t __s0_116 = __p0_116; \
+  float64x1_t __s1_116 = __p1_116; \
   float64x2_t __s2_116 = __p2_116; \
-  float64x2_t __ret_116; \
-  __ret_116 = vfmaq_laneq_f64(__s0_116, -__s1_116, __s2_116, __p3_116); \
+  float64x1_t __ret_116; \
+  __ret_116 = vfma_laneq_f64(__s0_116, -__s1_116, __s2_116, __p3_116); \
   __ret_116; \
 })
 #else
-#define vfmsq_laneq_f64(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
-  float64x2_t __s0_117 = __p0_117; \
-  float64x2_t __s1_117 = __p1_117; \
+#define vfms_laneq_f64(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
+  float64x1_t __s0_117 = __p0_117; \
+  float64x1_t __s1_117 = __p1_117; \
   float64x2_t __s2_117 = __p2_117; \
-  float64x2_t __rev0_117;  __rev0_117 = __builtin_shufflevector(__s0_117, __s0_117, 1, 0); \
-  float64x2_t __rev1_117;  __rev1_117 = __builtin_shufflevector(__s1_117, __s1_117, 1, 0); \
   float64x2_t __rev2_117;  __rev2_117 = __builtin_shufflevector(__s2_117, __s2_117, 1, 0); \
-  float64x2_t __ret_117; \
-  __ret_117 = __noswap_vfmaq_laneq_f64(__rev0_117, -__rev1_117, __rev2_117, __p3_117); \
-  __ret_117 = __builtin_shufflevector(__ret_117, __ret_117, 1, 0); \
+  float64x1_t __ret_117; \
+  __ret_117 = __noswap_vfma_laneq_f64(__s0_117, -__s1_117, __rev2_117, __p3_117); \
   __ret_117; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
-  float32x4_t __s0_118 = __p0_118; \
-  float32x4_t __s1_118 = __p1_118; \
+#define vfms_laneq_f32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
+  float32x2_t __s0_118 = __p0_118; \
+  float32x2_t __s1_118 = __p1_118; \
   float32x4_t __s2_118 = __p2_118; \
-  float32x4_t __ret_118; \
-  __ret_118 = vfmaq_laneq_f32(__s0_118, -__s1_118, __s2_118, __p3_118); \
+  float32x2_t __ret_118; \
+  __ret_118 = vfma_laneq_f32(__s0_118, -__s1_118, __s2_118, __p3_118); \
   __ret_118; \
 })
 #else
-#define vfmsq_laneq_f32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
-  float32x4_t __s0_119 = __p0_119; \
-  float32x4_t __s1_119 = __p1_119; \
+#define vfms_laneq_f32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
+  float32x2_t __s0_119 = __p0_119; \
+  float32x2_t __s1_119 = __p1_119; \
   float32x4_t __s2_119 = __p2_119; \
-  float32x4_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 3, 2, 1, 0); \
-  float32x4_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 3, 2, 1, 0); \
+  float32x2_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 1, 0); \
+  float32x2_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 1, 0); \
   float32x4_t __rev2_119;  __rev2_119 = __builtin_shufflevector(__s2_119, __s2_119, 3, 2, 1, 0); \
-  float32x4_t __ret_119; \
-  __ret_119 = __noswap_vfmaq_laneq_f32(__rev0_119, -__rev1_119, __rev2_119, __p3_119); \
-  __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 3, 2, 1, 0); \
+  float32x2_t __ret_119; \
+  __ret_119 = __noswap_vfma_laneq_f32(__rev0_119, -__rev1_119, __rev2_119, __p3_119); \
+  __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 1, 0); \
   __ret_119; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f64(__p0_120, __p1_120, __p2_120, __p3_120) __extension__ ({ \
-  float64x1_t __s0_120 = __p0_120; \
-  float64x1_t __s1_120 = __p1_120; \
-  float64x2_t __s2_120 = __p2_120; \
-  float64x1_t __ret_120; \
-  __ret_120 = vfma_laneq_f64(__s0_120, -__s1_120, __s2_120, __p3_120); \
-  __ret_120; \
-})
-#else
-#define vfms_laneq_f64(__p0_121, __p1_121, __p2_121, __p3_121) __extension__ ({ \
-  float64x1_t __s0_121 = __p0_121; \
-  float64x1_t __s1_121 = __p1_121; \
-  float64x2_t __s2_121 = __p2_121; \
-  float64x2_t __rev2_121;  __rev2_121 = __builtin_shufflevector(__s2_121, __s2_121, 1, 0); \
-  float64x1_t __ret_121; \
-  __ret_121 = __noswap_vfma_laneq_f64(__s0_121, -__s1_121, __rev2_121, __p3_121); \
-  __ret_121; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f32(__p0_122, __p1_122, __p2_122, __p3_122) __extension__ ({ \
-  float32x2_t __s0_122 = __p0_122; \
-  float32x2_t __s1_122 = __p1_122; \
-  float32x4_t __s2_122 = __p2_122; \
-  float32x2_t __ret_122; \
-  __ret_122 = vfma_laneq_f32(__s0_122, -__s1_122, __s2_122, __p3_122); \
-  __ret_122; \
-})
-#else
-#define vfms_laneq_f32(__p0_123, __p1_123, __p2_123, __p3_123) __extension__ ({ \
-  float32x2_t __s0_123 = __p0_123; \
-  float32x2_t __s1_123 = __p1_123; \
-  float32x4_t __s2_123 = __p2_123; \
-  float32x2_t __rev0_123;  __rev0_123 = __builtin_shufflevector(__s0_123, __s0_123, 1, 0); \
-  float32x2_t __rev1_123;  __rev1_123 = __builtin_shufflevector(__s1_123, __s1_123, 1, 0); \
-  float32x4_t __rev2_123;  __rev2_123 = __builtin_shufflevector(__s2_123, __s2_123, 3, 2, 1, 0); \
-  float32x2_t __ret_123; \
-  __ret_123 = __noswap_vfma_laneq_f32(__rev0_123, -__rev1_123, __rev2_123, __p3_123); \
-  __ret_123 = __builtin_shufflevector(__ret_123, __ret_123, 1, 0); \
-  __ret_123; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
   float64x2_t __ret;
   __ret = vfmaq_f64(__p0, -__p1, (float64x2_t) {__p2, __p2});
@@ -55078,20 +46574,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
   float64x1_t __ret;
   __ret = vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
   return __ret;
 }
-#else
-__ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
   float32x2_t __ret;
@@ -55144,28 +46631,12 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64_t __ret; \
   __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -55212,28 +46683,12 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vget_lane_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64_t __ret; \
   __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
   __ret; \
 })
-#else
-#define vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
   poly64x1_t __ret;
@@ -55264,20 +46719,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64(__p0) __extension__ ({ \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64(__p0) __extension__ ({ \
   poly64x2_t __ret; \
@@ -55308,34 +46754,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64(__p0) __extension__ ({ \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_p64(__p0) __extension__ ({ \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
   __ret; \
 })
-#else
-#define vld1_dup_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_dup_p64(__p0) __extension__ ({ \
   poly64x2_t __ret; \
@@ -55366,36 +46794,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_dup_f64(__p0) __extension__ ({ \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
   __ret; \
 })
-#else
-#define vld1_dup_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   poly64x1_t __ret; \
   __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -55432,36 +46841,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   float64x1_t __ret; \
   __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x2(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x2(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x2(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55496,34 +46886,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x2(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x2(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x3(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x3(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x3(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -55560,34 +46932,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x3(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x3(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld1_p64_x4(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld1_p64_x4(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld1q_p64_x4(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -55626,34 +46980,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld1_f64_x4(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld1_f64_x4(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_p64(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld2_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_p64(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55722,34 +47058,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_f64(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld2_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_p64(__p0) __extension__ ({ \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld2_dup_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_dup_p64(__p0) __extension__ ({ \
   poly64x2x2_t __ret; \
@@ -55784,36 +47102,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_dup_f64(__p0) __extension__ ({ \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld2_dup_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   poly64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
   __ret; \
 })
-#else
-#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x2_t __s1 = __p1; \
@@ -55968,68 +47267,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   uint64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
   __ret; \
 })
-#else
-#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   float64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
   __ret; \
 })
-#else
-#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   int64x1x2_t __ret; \
   __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
   __ret; \
 })
-#else
-#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_p64(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld3_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_p64(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -56102,34 +47362,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_f64(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld3_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_p64(__p0) __extension__ ({ \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld3_dup_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_dup_p64(__p0) __extension__ ({ \
   poly64x2x3_t __ret; \
@@ -56166,36 +47408,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_dup_f64(__p0) __extension__ ({ \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld3_dup_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   poly64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
   __ret; \
 })
-#else
-#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x3_t __s1 = __p1; \
@@ -56364,68 +47587,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   uint64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
   __ret; \
 })
-#else
-#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   float64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
   __ret; \
 })
-#else
-#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   int64x1x3_t __ret; \
   __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
   __ret; \
 })
-#else
-#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_p64(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld4_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_p64(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -56502,34 +47686,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_f64(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld4_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_p64(__p0) __extension__ ({ \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
   __ret; \
 })
-#else
-#define vld4_dup_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_dup_p64(__p0) __extension__ ({ \
   poly64x2x4_t __ret; \
@@ -56568,36 +47734,17 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_dup_f64(__p0) __extension__ ({ \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
   __ret; \
 })
-#else
-#define vld4_dup_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   poly64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
   __ret; \
 })
-#else
-#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x4_t __s1 = __p1; \
@@ -56780,68 +47927,29 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   uint64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
   __ret; \
 })
-#else
-#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   float64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
   __ret; \
 })
-#else
-#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   int64x1x4_t __ret; \
   __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
   __ret; \
 })
-#else
-#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vldrq_p128(__p0) __extension__ ({ \
   poly128_t __ret; \
   __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
   __ret; \
 })
-#else
-#define vldrq_p128(__p0) __extension__ ({ \
-  poly128_t __ret; \
-  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -56859,20 +47967,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
   float64_t __ret;
@@ -57160,20 +48259,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64_t vminnmvq_f64(float64x2_t __p0) {
   float64_t __ret;
@@ -57462,20 +48552,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = __p0 + __p1 * __p2;
   return __ret;
 }
-#else
-__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
   uint32x4_t __s0 = __p0; \
@@ -58039,20 +49120,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
   float64x1_t __ret;
   __ret = __p0 - __p1 * __p2;
   return __ret;
 }
-#else
-__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
   uint32x4_t __s0 = __p0; \
@@ -58598,20 +49670,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly64x1_t vmov_n_p64(poly64_t __p0) {
   poly64x1_t __ret;
   __ret = (poly64x1_t) {__p0};
   return __ret;
 }
-#else
-__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
   poly64x2_t __ret;
@@ -58642,162 +49705,153 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmov_n_f64(float64_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) {__p0};
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_120) {
+  uint16x8_t __ret_120;
+  uint8x8_t __a1_120 = vget_high_u8(__p0_120);
+  __ret_120 = (uint16x8_t)(vshll_n_u8(__a1_120, 0));
+  return __ret_120;
+}
 #else
-__ai float64x1_t vmov_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
+__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_121) {
+  uint8x16_t __rev0_121;  __rev0_121 = __builtin_shufflevector(__p0_121, __p0_121, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint16x8_t __ret_121;
+  uint8x8_t __a1_121 = __noswap_vget_high_u8(__rev0_121);
+  __ret_121 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_121, 0));
+  __ret_121 = __builtin_shufflevector(__ret_121, __ret_121, 7, 6, 5, 4, 3, 2, 1, 0);
+  return __ret_121;
+}
+__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_122) {
+  uint16x8_t __ret_122;
+  uint8x8_t __a1_122 = __noswap_vget_high_u8(__p0_122);
+  __ret_122 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_122, 0));
+  return __ret_122;
 }
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_124) {
-  uint16x8_t __ret_124;
-  uint8x8_t __a1_124 = vget_high_u8(__p0_124);
-  __ret_124 = (uint16x8_t)(vshll_n_u8(__a1_124, 0));
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_123) {
+  uint64x2_t __ret_123;
+  uint32x2_t __a1_123 = vget_high_u32(__p0_123);
+  __ret_123 = (uint64x2_t)(vshll_n_u32(__a1_123, 0));
+  return __ret_123;
+}
+#else
+__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_124) {
+  uint32x4_t __rev0_124;  __rev0_124 = __builtin_shufflevector(__p0_124, __p0_124, 3, 2, 1, 0);
+  uint64x2_t __ret_124;
+  uint32x2_t __a1_124 = __noswap_vget_high_u32(__rev0_124);
+  __ret_124 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_124, 0));
+  __ret_124 = __builtin_shufflevector(__ret_124, __ret_124, 1, 0);
   return __ret_124;
 }
-#else
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_125) {
-  uint8x16_t __rev0_125;  __rev0_125 = __builtin_shufflevector(__p0_125, __p0_125, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret_125;
-  uint8x8_t __a1_125 = __noswap_vget_high_u8(__rev0_125);
-  __ret_125 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_125, 0));
-  __ret_125 = __builtin_shufflevector(__ret_125, __ret_125, 7, 6, 5, 4, 3, 2, 1, 0);
+__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_125) {
+  uint64x2_t __ret_125;
+  uint32x2_t __a1_125 = __noswap_vget_high_u32(__p0_125);
+  __ret_125 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_125, 0));
   return __ret_125;
 }
-__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_126) {
-  uint16x8_t __ret_126;
-  uint8x8_t __a1_126 = __noswap_vget_high_u8(__p0_126);
-  __ret_126 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_126, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_126) {
+  uint32x4_t __ret_126;
+  uint16x4_t __a1_126 = vget_high_u16(__p0_126);
+  __ret_126 = (uint32x4_t)(vshll_n_u16(__a1_126, 0));
   return __ret_126;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_127) {
-  uint64x2_t __ret_127;
-  uint32x2_t __a1_127 = vget_high_u32(__p0_127);
-  __ret_127 = (uint64x2_t)(vshll_n_u32(__a1_127, 0));
+#else
+__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_127) {
+  uint16x8_t __rev0_127;  __rev0_127 = __builtin_shufflevector(__p0_127, __p0_127, 7, 6, 5, 4, 3, 2, 1, 0);
+  uint32x4_t __ret_127;
+  uint16x4_t __a1_127 = __noswap_vget_high_u16(__rev0_127);
+  __ret_127 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_127, 0));
+  __ret_127 = __builtin_shufflevector(__ret_127, __ret_127, 3, 2, 1, 0);
   return __ret_127;
 }
-#else
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_128) {
-  uint32x4_t __rev0_128;  __rev0_128 = __builtin_shufflevector(__p0_128, __p0_128, 3, 2, 1, 0);
-  uint64x2_t __ret_128;
-  uint32x2_t __a1_128 = __noswap_vget_high_u32(__rev0_128);
-  __ret_128 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_128, 0));
-  __ret_128 = __builtin_shufflevector(__ret_128, __ret_128, 1, 0);
+__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_128) {
+  uint32x4_t __ret_128;
+  uint16x4_t __a1_128 = __noswap_vget_high_u16(__p0_128);
+  __ret_128 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_128, 0));
   return __ret_128;
 }
-__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_129) {
-  uint64x2_t __ret_129;
-  uint32x2_t __a1_129 = __noswap_vget_high_u32(__p0_129);
-  __ret_129 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_129, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_129) {
+  int16x8_t __ret_129;
+  int8x8_t __a1_129 = vget_high_s8(__p0_129);
+  __ret_129 = (int16x8_t)(vshll_n_s8(__a1_129, 0));
   return __ret_129;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_130) {
-  uint32x4_t __ret_130;
-  uint16x4_t __a1_130 = vget_high_u16(__p0_130);
-  __ret_130 = (uint32x4_t)(vshll_n_u16(__a1_130, 0));
+#else
+__ai int16x8_t vmovl_high_s8(int8x16_t __p0_130) {
+  int8x16_t __rev0_130;  __rev0_130 = __builtin_shufflevector(__p0_130, __p0_130, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
+  int16x8_t __ret_130;
+  int8x8_t __a1_130 = __noswap_vget_high_s8(__rev0_130);
+  __ret_130 = (int16x8_t)(__noswap_vshll_n_s8(__a1_130, 0));
+  __ret_130 = __builtin_shufflevector(__ret_130, __ret_130, 7, 6, 5, 4, 3, 2, 1, 0);
   return __ret_130;
 }
-#else
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_131) {
-  uint16x8_t __rev0_131;  __rev0_131 = __builtin_shufflevector(__p0_131, __p0_131, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret_131;
-  uint16x4_t __a1_131 = __noswap_vget_high_u16(__rev0_131);
-  __ret_131 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_131, 0));
-  __ret_131 = __builtin_shufflevector(__ret_131, __ret_131, 3, 2, 1, 0);
+__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_131) {
+  int16x8_t __ret_131;
+  int8x8_t __a1_131 = __noswap_vget_high_s8(__p0_131);
+  __ret_131 = (int16x8_t)(__noswap_vshll_n_s8(__a1_131, 0));
   return __ret_131;
 }
-__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_132) {
-  uint32x4_t __ret_132;
-  uint16x4_t __a1_132 = __noswap_vget_high_u16(__p0_132);
-  __ret_132 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_132, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_132) {
+  int64x2_t __ret_132;
+  int32x2_t __a1_132 = vget_high_s32(__p0_132);
+  __ret_132 = (int64x2_t)(vshll_n_s32(__a1_132, 0));
   return __ret_132;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_133) {
-  int16x8_t __ret_133;
-  int8x8_t __a1_133 = vget_high_s8(__p0_133);
-  __ret_133 = (int16x8_t)(vshll_n_s8(__a1_133, 0));
+#else
+__ai int64x2_t vmovl_high_s32(int32x4_t __p0_133) {
+  int32x4_t __rev0_133;  __rev0_133 = __builtin_shufflevector(__p0_133, __p0_133, 3, 2, 1, 0);
+  int64x2_t __ret_133;
+  int32x2_t __a1_133 = __noswap_vget_high_s32(__rev0_133);
+  __ret_133 = (int64x2_t)(__noswap_vshll_n_s32(__a1_133, 0));
+  __ret_133 = __builtin_shufflevector(__ret_133, __ret_133, 1, 0);
   return __ret_133;
 }
-#else
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_134) {
-  int8x16_t __rev0_134;  __rev0_134 = __builtin_shufflevector(__p0_134, __p0_134, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret_134;
-  int8x8_t __a1_134 = __noswap_vget_high_s8(__rev0_134);
-  __ret_134 = (int16x8_t)(__noswap_vshll_n_s8(__a1_134, 0));
-  __ret_134 = __builtin_shufflevector(__ret_134, __ret_134, 7, 6, 5, 4, 3, 2, 1, 0);
+__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_134) {
+  int64x2_t __ret_134;
+  int32x2_t __a1_134 = __noswap_vget_high_s32(__p0_134);
+  __ret_134 = (int64x2_t)(__noswap_vshll_n_s32(__a1_134, 0));
   return __ret_134;
 }
-__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_135) {
-  int16x8_t __ret_135;
-  int8x8_t __a1_135 = __noswap_vget_high_s8(__p0_135);
-  __ret_135 = (int16x8_t)(__noswap_vshll_n_s8(__a1_135, 0));
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_135) {
+  int32x4_t __ret_135;
+  int16x4_t __a1_135 = vget_high_s16(__p0_135);
+  __ret_135 = (int32x4_t)(vshll_n_s16(__a1_135, 0));
   return __ret_135;
 }
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_136) {
-  int64x2_t __ret_136;
-  int32x2_t __a1_136 = vget_high_s32(__p0_136);
-  __ret_136 = (int64x2_t)(vshll_n_s32(__a1_136, 0));
+#else
+__ai int32x4_t vmovl_high_s16(int16x8_t __p0_136) {
+  int16x8_t __rev0_136;  __rev0_136 = __builtin_shufflevector(__p0_136, __p0_136, 7, 6, 5, 4, 3, 2, 1, 0);
+  int32x4_t __ret_136;
+  int16x4_t __a1_136 = __noswap_vget_high_s16(__rev0_136);
+  __ret_136 = (int32x4_t)(__noswap_vshll_n_s16(__a1_136, 0));
+  __ret_136 = __builtin_shufflevector(__ret_136, __ret_136, 3, 2, 1, 0);
   return __ret_136;
 }
-#else
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_137) {
-  int32x4_t __rev0_137;  __rev0_137 = __builtin_shufflevector(__p0_137, __p0_137, 3, 2, 1, 0);
-  int64x2_t __ret_137;
-  int32x2_t __a1_137 = __noswap_vget_high_s32(__rev0_137);
-  __ret_137 = (int64x2_t)(__noswap_vshll_n_s32(__a1_137, 0));
-  __ret_137 = __builtin_shufflevector(__ret_137, __ret_137, 1, 0);
+__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_137) {
+  int32x4_t __ret_137;
+  int16x4_t __a1_137 = __noswap_vget_high_s16(__p0_137);
+  __ret_137 = (int32x4_t)(__noswap_vshll_n_s16(__a1_137, 0));
   return __ret_137;
 }
-__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_138) {
-  int64x2_t __ret_138;
-  int32x2_t __a1_138 = __noswap_vget_high_s32(__p0_138);
-  __ret_138 = (int64x2_t)(__noswap_vshll_n_s32(__a1_138, 0));
-  return __ret_138;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_139) {
-  int32x4_t __ret_139;
-  int16x4_t __a1_139 = vget_high_s16(__p0_139);
-  __ret_139 = (int32x4_t)(vshll_n_s16(__a1_139, 0));
-  return __ret_139;
-}
-#else
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_140) {
-  int16x8_t __rev0_140;  __rev0_140 = __builtin_shufflevector(__p0_140, __p0_140, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret_140;
-  int16x4_t __a1_140 = __noswap_vget_high_s16(__rev0_140);
-  __ret_140 = (int32x4_t)(__noswap_vshll_n_s16(__a1_140, 0));
-  __ret_140 = __builtin_shufflevector(__ret_140, __ret_140, 3, 2, 1, 0);
-  return __ret_140;
-}
-__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_141) {
-  int32x4_t __ret_141;
-  int16x4_t __a1_141 = __noswap_vget_high_s16(__p0_141);
-  __ret_141 = (int32x4_t)(__noswap_vshll_n_s16(__a1_141, 0));
-  return __ret_141;
-}
 #endif
 
 #ifdef __LITTLE_ENDIAN__
@@ -58919,58 +49973,37 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 * __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#endif
-
+#define vmuld_lane_f64(__p0_138, __p1_138, __p2_138) __extension__ ({ \
+  float64_t __s0_138 = __p0_138; \
+  float64x1_t __s1_138 = __p1_138; \
+  float64_t __ret_138; \
+  __ret_138 = __s0_138 * vget_lane_f64(__s1_138, __p2_138); \
+  __ret_138; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmuld_lane_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
-  float64_t __s0_142 = __p0_142; \
-  float64x1_t __s1_142 = __p1_142; \
-  float64_t __ret_142; \
-  __ret_142 = __s0_142 * vget_lane_f64(__s1_142, __p2_142); \
-  __ret_142; \
+#define vmuls_lane_f32(__p0_139, __p1_139, __p2_139) __extension__ ({ \
+  float32_t __s0_139 = __p0_139; \
+  float32x2_t __s1_139 = __p1_139; \
+  float32_t __ret_139; \
+  __ret_139 = __s0_139 * vget_lane_f32(__s1_139, __p2_139); \
+  __ret_139; \
 })
 #else
-#define vmuld_lane_f64(__p0_143, __p1_143, __p2_143) __extension__ ({ \
-  float64_t __s0_143 = __p0_143; \
-  float64x1_t __s1_143 = __p1_143; \
-  float64_t __ret_143; \
-  __ret_143 = __s0_143 * __noswap_vget_lane_f64(__s1_143, __p2_143); \
-  __ret_143; \
+#define vmuls_lane_f32(__p0_140, __p1_140, __p2_140) __extension__ ({ \
+  float32_t __s0_140 = __p0_140; \
+  float32x2_t __s1_140 = __p1_140; \
+  float32x2_t __rev1_140;  __rev1_140 = __builtin_shufflevector(__s1_140, __s1_140, 1, 0); \
+  float32_t __ret_140; \
+  __ret_140 = __s0_140 * __noswap_vget_lane_f32(__rev1_140, __p2_140); \
+  __ret_140; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vmuls_lane_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
-  float32_t __s0_144 = __p0_144; \
-  float32x2_t __s1_144 = __p1_144; \
-  float32_t __ret_144; \
-  __ret_144 = __s0_144 * vget_lane_f32(__s1_144, __p2_144); \
-  __ret_144; \
-})
-#else
-#define vmuls_lane_f32(__p0_145, __p1_145, __p2_145) __extension__ ({ \
-  float32_t __s0_145 = __p0_145; \
-  float32x2_t __s1_145 = __p1_145; \
-  float32x2_t __rev1_145;  __rev1_145 = __builtin_shufflevector(__s1_145, __s1_145, 1, 0); \
-  float32_t __ret_145; \
-  __ret_145 = __s0_145 * __noswap_vget_lane_f32(__rev1_145, __p2_145); \
-  __ret_145; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -58978,16 +50011,6 @@
   __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
   __ret; \
 })
-#else
-#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x2_t __s0 = __p0; \
@@ -59009,40 +50032,40 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmuld_laneq_f64(__p0_146, __p1_146, __p2_146) __extension__ ({ \
-  float64_t __s0_146 = __p0_146; \
-  float64x2_t __s1_146 = __p1_146; \
-  float64_t __ret_146; \
-  __ret_146 = __s0_146 * vgetq_lane_f64(__s1_146, __p2_146); \
-  __ret_146; \
+#define vmuld_laneq_f64(__p0_141, __p1_141, __p2_141) __extension__ ({ \
+  float64_t __s0_141 = __p0_141; \
+  float64x2_t __s1_141 = __p1_141; \
+  float64_t __ret_141; \
+  __ret_141 = __s0_141 * vgetq_lane_f64(__s1_141, __p2_141); \
+  __ret_141; \
 })
 #else
-#define vmuld_laneq_f64(__p0_147, __p1_147, __p2_147) __extension__ ({ \
-  float64_t __s0_147 = __p0_147; \
-  float64x2_t __s1_147 = __p1_147; \
-  float64x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
-  float64_t __ret_147; \
-  __ret_147 = __s0_147 * __noswap_vgetq_lane_f64(__rev1_147, __p2_147); \
-  __ret_147; \
+#define vmuld_laneq_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
+  float64_t __s0_142 = __p0_142; \
+  float64x2_t __s1_142 = __p1_142; \
+  float64x2_t __rev1_142;  __rev1_142 = __builtin_shufflevector(__s1_142, __s1_142, 1, 0); \
+  float64_t __ret_142; \
+  __ret_142 = __s0_142 * __noswap_vgetq_lane_f64(__rev1_142, __p2_142); \
+  __ret_142; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmuls_laneq_f32(__p0_148, __p1_148, __p2_148) __extension__ ({ \
-  float32_t __s0_148 = __p0_148; \
-  float32x4_t __s1_148 = __p1_148; \
-  float32_t __ret_148; \
-  __ret_148 = __s0_148 * vgetq_lane_f32(__s1_148, __p2_148); \
-  __ret_148; \
+#define vmuls_laneq_f32(__p0_143, __p1_143, __p2_143) __extension__ ({ \
+  float32_t __s0_143 = __p0_143; \
+  float32x4_t __s1_143 = __p1_143; \
+  float32_t __ret_143; \
+  __ret_143 = __s0_143 * vgetq_lane_f32(__s1_143, __p2_143); \
+  __ret_143; \
 })
 #else
-#define vmuls_laneq_f32(__p0_149, __p1_149, __p2_149) __extension__ ({ \
-  float32_t __s0_149 = __p0_149; \
-  float32x4_t __s1_149 = __p1_149; \
-  float32x4_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 3, 2, 1, 0); \
-  float32_t __ret_149; \
-  __ret_149 = __s0_149 * __noswap_vgetq_lane_f32(__rev1_149, __p2_149); \
-  __ret_149; \
+#define vmuls_laneq_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
+  float32_t __s0_144 = __p0_144; \
+  float32x4_t __s1_144 = __p1_144; \
+  float32x4_t __rev1_144;  __rev1_144 = __builtin_shufflevector(__s1_144, __s1_144, 3, 2, 1, 0); \
+  float32_t __ret_144; \
+  __ret_144 = __s0_144 * __noswap_vgetq_lane_f32(__rev1_144, __p2_144); \
+  __ret_144; \
 })
 #endif
 
@@ -59296,20 +50319,11 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
   return __ret;
 }
-#else
-__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
   float64x2_t __ret;
@@ -59326,25 +50340,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
   poly128_t __ret;
   __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
   return __ret;
 }
-#else
-__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-__ai poly128_t __noswap_vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
   poly16x8_t __ret;
@@ -59475,7 +50475,7 @@
   poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
   poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
   poly128_t __ret;
-  __ret = __noswap_vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
+  __ret = vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
   return __ret;
 }
 #endif
@@ -59840,20 +50840,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
   float32x2_t __ret;
@@ -59876,78 +50867,39 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-__ai float64_t __noswap_vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-__ai float32_t __noswap_vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
+#define vmulxd_lane_f64(__p0_145, __p1_145, __p2_145) __extension__ ({ \
+  float64_t __s0_145 = __p0_145; \
+  float64x1_t __s1_145 = __p1_145; \
+  float64_t __ret_145; \
+  __ret_145 = vmulxd_f64(__s0_145, vget_lane_f64(__s1_145, __p2_145)); \
+  __ret_145; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmulxd_lane_f64(__p0_150, __p1_150, __p2_150) __extension__ ({ \
-  float64_t __s0_150 = __p0_150; \
-  float64x1_t __s1_150 = __p1_150; \
-  float64_t __ret_150; \
-  __ret_150 = vmulxd_f64(__s0_150, vget_lane_f64(__s1_150, __p2_150)); \
-  __ret_150; \
+#define vmulxs_lane_f32(__p0_146, __p1_146, __p2_146) __extension__ ({ \
+  float32_t __s0_146 = __p0_146; \
+  float32x2_t __s1_146 = __p1_146; \
+  float32_t __ret_146; \
+  __ret_146 = vmulxs_f32(__s0_146, vget_lane_f32(__s1_146, __p2_146)); \
+  __ret_146; \
 })
 #else
-#define vmulxd_lane_f64(__p0_151, __p1_151, __p2_151) __extension__ ({ \
-  float64_t __s0_151 = __p0_151; \
-  float64x1_t __s1_151 = __p1_151; \
-  float64_t __ret_151; \
-  __ret_151 = __noswap_vmulxd_f64(__s0_151, __noswap_vget_lane_f64(__s1_151, __p2_151)); \
-  __ret_151; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxs_lane_f32(__p0_152, __p1_152, __p2_152) __extension__ ({ \
-  float32_t __s0_152 = __p0_152; \
-  float32x2_t __s1_152 = __p1_152; \
-  float32_t __ret_152; \
-  __ret_152 = vmulxs_f32(__s0_152, vget_lane_f32(__s1_152, __p2_152)); \
-  __ret_152; \
-})
-#else
-#define vmulxs_lane_f32(__p0_153, __p1_153, __p2_153) __extension__ ({ \
-  float32_t __s0_153 = __p0_153; \
-  float32x2_t __s1_153 = __p1_153; \
-  float32x2_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 1, 0); \
-  float32_t __ret_153; \
-  __ret_153 = __noswap_vmulxs_f32(__s0_153, __noswap_vget_lane_f32(__rev1_153, __p2_153)); \
-  __ret_153; \
+#define vmulxs_lane_f32(__p0_147, __p1_147, __p2_147) __extension__ ({ \
+  float32_t __s0_147 = __p0_147; \
+  float32x2_t __s1_147 = __p1_147; \
+  float32x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
+  float32_t __ret_147; \
+  __ret_147 = vmulxs_f32(__s0_147, __noswap_vget_lane_f32(__rev1_147, __p2_147)); \
+  __ret_147; \
 })
 #endif
 
@@ -60014,40 +50966,40 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulxd_laneq_f64(__p0_154, __p1_154, __p2_154) __extension__ ({ \
-  float64_t __s0_154 = __p0_154; \
-  float64x2_t __s1_154 = __p1_154; \
-  float64_t __ret_154; \
-  __ret_154 = vmulxd_f64(__s0_154, vgetq_lane_f64(__s1_154, __p2_154)); \
-  __ret_154; \
+#define vmulxd_laneq_f64(__p0_148, __p1_148, __p2_148) __extension__ ({ \
+  float64_t __s0_148 = __p0_148; \
+  float64x2_t __s1_148 = __p1_148; \
+  float64_t __ret_148; \
+  __ret_148 = vmulxd_f64(__s0_148, vgetq_lane_f64(__s1_148, __p2_148)); \
+  __ret_148; \
 })
 #else
-#define vmulxd_laneq_f64(__p0_155, __p1_155, __p2_155) __extension__ ({ \
-  float64_t __s0_155 = __p0_155; \
-  float64x2_t __s1_155 = __p1_155; \
-  float64x2_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 1, 0); \
-  float64_t __ret_155; \
-  __ret_155 = __noswap_vmulxd_f64(__s0_155, __noswap_vgetq_lane_f64(__rev1_155, __p2_155)); \
-  __ret_155; \
+#define vmulxd_laneq_f64(__p0_149, __p1_149, __p2_149) __extension__ ({ \
+  float64_t __s0_149 = __p0_149; \
+  float64x2_t __s1_149 = __p1_149; \
+  float64x2_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 1, 0); \
+  float64_t __ret_149; \
+  __ret_149 = vmulxd_f64(__s0_149, __noswap_vgetq_lane_f64(__rev1_149, __p2_149)); \
+  __ret_149; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulxs_laneq_f32(__p0_156, __p1_156, __p2_156) __extension__ ({ \
-  float32_t __s0_156 = __p0_156; \
-  float32x4_t __s1_156 = __p1_156; \
-  float32_t __ret_156; \
-  __ret_156 = vmulxs_f32(__s0_156, vgetq_lane_f32(__s1_156, __p2_156)); \
-  __ret_156; \
+#define vmulxs_laneq_f32(__p0_150, __p1_150, __p2_150) __extension__ ({ \
+  float32_t __s0_150 = __p0_150; \
+  float32x4_t __s1_150 = __p1_150; \
+  float32_t __ret_150; \
+  __ret_150 = vmulxs_f32(__s0_150, vgetq_lane_f32(__s1_150, __p2_150)); \
+  __ret_150; \
 })
 #else
-#define vmulxs_laneq_f32(__p0_157, __p1_157, __p2_157) __extension__ ({ \
-  float32_t __s0_157 = __p0_157; \
-  float32x4_t __s1_157 = __p1_157; \
-  float32x4_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 3, 2, 1, 0); \
-  float32_t __ret_157; \
-  __ret_157 = __noswap_vmulxs_f32(__s0_157, __noswap_vgetq_lane_f32(__rev1_157, __p2_157)); \
-  __ret_157; \
+#define vmulxs_laneq_f32(__p0_151, __p1_151, __p2_151) __extension__ ({ \
+  float32_t __s0_151 = __p0_151; \
+  float32x4_t __s1_151 = __p1_151; \
+  float32x4_t __rev1_151;  __rev1_151 = __builtin_shufflevector(__s1_151, __s1_151, 3, 2, 1, 0); \
+  float32_t __ret_151; \
+  __ret_151 = vmulxs_f32(__s0_151, __noswap_vgetq_lane_f32(__rev1_151, __p2_151)); \
+  __ret_151; \
 })
 #endif
 
@@ -60146,48 +51098,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vneg_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = -__p0;
   return __ret;
 }
-#else
-__ai float64x1_t vneg_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vneg_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = -__p0;
   return __ret;
 }
-#else
-__ai int64x1_t vneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vnegd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
   uint8x16_t __ret;
@@ -60928,226 +51853,81 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqabs_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqabsb_s8(int8_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqabsb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqabss_s32(int32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqabss_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqabsd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vqabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqabsh_s16(int16_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqabsh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
   int64x2_t __ret;
@@ -61446,34 +52226,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
   return __ret;
 }
-#else
-__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
   int64x2_t __ret;
@@ -61772,121 +52534,93 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_lane_s32(__p0_152, __p1_152, __p2_152) __extension__ ({ \
+  int32_t __s0_152 = __p0_152; \
+  int32x2_t __s1_152 = __p1_152; \
+  int32_t __ret_152; \
+  __ret_152 = vqdmulhs_s32(__s0_152, vget_lane_s32(__s1_152, __p2_152)); \
+  __ret_152; \
+})
 #else
-__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqdmulhs_lane_s32(__p0_153, __p1_153, __p2_153) __extension__ ({ \
+  int32_t __s0_153 = __p0_153; \
+  int32x2_t __s1_153 = __p1_153; \
+  int32x2_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 1, 0); \
+  int32_t __ret_153; \
+  __ret_153 = vqdmulhs_s32(__s0_153, __noswap_vget_lane_s32(__rev1_153, __p2_153)); \
+  __ret_153; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_lane_s32(__p0_158, __p1_158, __p2_158) __extension__ ({ \
-  int32_t __s0_158 = __p0_158; \
-  int32x2_t __s1_158 = __p1_158; \
-  int32_t __ret_158; \
-  __ret_158 = vqdmulhs_s32(__s0_158, vget_lane_s32(__s1_158, __p2_158)); \
+#define vqdmulhh_lane_s16(__p0_154, __p1_154, __p2_154) __extension__ ({ \
+  int16_t __s0_154 = __p0_154; \
+  int16x4_t __s1_154 = __p1_154; \
+  int16_t __ret_154; \
+  __ret_154 = vqdmulhh_s16(__s0_154, vget_lane_s16(__s1_154, __p2_154)); \
+  __ret_154; \
+})
+#else
+#define vqdmulhh_lane_s16(__p0_155, __p1_155, __p2_155) __extension__ ({ \
+  int16_t __s0_155 = __p0_155; \
+  int16x4_t __s1_155 = __p1_155; \
+  int16x4_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 3, 2, 1, 0); \
+  int16_t __ret_155; \
+  __ret_155 = vqdmulhh_s16(__s0_155, __noswap_vget_lane_s16(__rev1_155, __p2_155)); \
+  __ret_155; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhs_laneq_s32(__p0_156, __p1_156, __p2_156) __extension__ ({ \
+  int32_t __s0_156 = __p0_156; \
+  int32x4_t __s1_156 = __p1_156; \
+  int32_t __ret_156; \
+  __ret_156 = vqdmulhs_s32(__s0_156, vgetq_lane_s32(__s1_156, __p2_156)); \
+  __ret_156; \
+})
+#else
+#define vqdmulhs_laneq_s32(__p0_157, __p1_157, __p2_157) __extension__ ({ \
+  int32_t __s0_157 = __p0_157; \
+  int32x4_t __s1_157 = __p1_157; \
+  int32x4_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 3, 2, 1, 0); \
+  int32_t __ret_157; \
+  __ret_157 = vqdmulhs_s32(__s0_157, __noswap_vgetq_lane_s32(__rev1_157, __p2_157)); \
+  __ret_157; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulhh_laneq_s16(__p0_158, __p1_158, __p2_158) __extension__ ({ \
+  int16_t __s0_158 = __p0_158; \
+  int16x8_t __s1_158 = __p1_158; \
+  int16_t __ret_158; \
+  __ret_158 = vqdmulhh_s16(__s0_158, vgetq_lane_s16(__s1_158, __p2_158)); \
   __ret_158; \
 })
 #else
-#define vqdmulhs_lane_s32(__p0_159, __p1_159, __p2_159) __extension__ ({ \
-  int32_t __s0_159 = __p0_159; \
-  int32x2_t __s1_159 = __p1_159; \
-  int32x2_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 1, 0); \
-  int32_t __ret_159; \
-  __ret_159 = __noswap_vqdmulhs_s32(__s0_159, __noswap_vget_lane_s32(__rev1_159, __p2_159)); \
+#define vqdmulhh_laneq_s16(__p0_159, __p1_159, __p2_159) __extension__ ({ \
+  int16_t __s0_159 = __p0_159; \
+  int16x8_t __s1_159 = __p1_159; \
+  int16x8_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_159; \
+  __ret_159 = vqdmulhh_s16(__s0_159, __noswap_vgetq_lane_s16(__rev1_159, __p2_159)); \
   __ret_159; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_lane_s16(__p0_160, __p1_160, __p2_160) __extension__ ({ \
-  int16_t __s0_160 = __p0_160; \
-  int16x4_t __s1_160 = __p1_160; \
-  int16_t __ret_160; \
-  __ret_160 = vqdmulhh_s16(__s0_160, vget_lane_s16(__s1_160, __p2_160)); \
-  __ret_160; \
-})
-#else
-#define vqdmulhh_lane_s16(__p0_161, __p1_161, __p2_161) __extension__ ({ \
-  int16_t __s0_161 = __p0_161; \
-  int16x4_t __s1_161 = __p1_161; \
-  int16x4_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 3, 2, 1, 0); \
-  int16_t __ret_161; \
-  __ret_161 = __noswap_vqdmulhh_s16(__s0_161, __noswap_vget_lane_s16(__rev1_161, __p2_161)); \
-  __ret_161; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_laneq_s32(__p0_162, __p1_162, __p2_162) __extension__ ({ \
-  int32_t __s0_162 = __p0_162; \
-  int32x4_t __s1_162 = __p1_162; \
-  int32_t __ret_162; \
-  __ret_162 = vqdmulhs_s32(__s0_162, vgetq_lane_s32(__s1_162, __p2_162)); \
-  __ret_162; \
-})
-#else
-#define vqdmulhs_laneq_s32(__p0_163, __p1_163, __p2_163) __extension__ ({ \
-  int32_t __s0_163 = __p0_163; \
-  int32x4_t __s1_163 = __p1_163; \
-  int32x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
-  int32_t __ret_163; \
-  __ret_163 = __noswap_vqdmulhs_s32(__s0_163, __noswap_vgetq_lane_s32(__rev1_163, __p2_163)); \
-  __ret_163; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_laneq_s16(__p0_164, __p1_164, __p2_164) __extension__ ({ \
-  int16_t __s0_164 = __p0_164; \
-  int16x8_t __s1_164 = __p1_164; \
-  int16_t __ret_164; \
-  __ret_164 = vqdmulhh_s16(__s0_164, vgetq_lane_s16(__s1_164, __p2_164)); \
-  __ret_164; \
-})
-#else
-#define vqdmulhh_laneq_s16(__p0_165, __p1_165, __p2_165) __extension__ ({ \
-  int16_t __s0_165 = __p0_165; \
-  int16x8_t __s1_165 = __p1_165; \
-  int16x8_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_165; \
-  __ret_165 = __noswap_vqdmulhh_s16(__s0_165, __noswap_vgetq_lane_s16(__rev1_165, __p2_165)); \
-  __ret_165; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x4_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -61970,44 +52704,16 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-__ai int64_t __noswap_vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
   int64x2_t __ret;
@@ -62159,82 +52865,82 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmulls_lane_s32(__p0_166, __p1_166, __p2_166) __extension__ ({ \
-  int32_t __s0_166 = __p0_166; \
-  int32x2_t __s1_166 = __p1_166; \
-  int64_t __ret_166; \
-  __ret_166 = vqdmulls_s32(__s0_166, vget_lane_s32(__s1_166, __p2_166)); \
+#define vqdmulls_lane_s32(__p0_160, __p1_160, __p2_160) __extension__ ({ \
+  int32_t __s0_160 = __p0_160; \
+  int32x2_t __s1_160 = __p1_160; \
+  int64_t __ret_160; \
+  __ret_160 = vqdmulls_s32(__s0_160, vget_lane_s32(__s1_160, __p2_160)); \
+  __ret_160; \
+})
+#else
+#define vqdmulls_lane_s32(__p0_161, __p1_161, __p2_161) __extension__ ({ \
+  int32_t __s0_161 = __p0_161; \
+  int32x2_t __s1_161 = __p1_161; \
+  int32x2_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 1, 0); \
+  int64_t __ret_161; \
+  __ret_161 = vqdmulls_s32(__s0_161, __noswap_vget_lane_s32(__rev1_161, __p2_161)); \
+  __ret_161; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_lane_s16(__p0_162, __p1_162, __p2_162) __extension__ ({ \
+  int16_t __s0_162 = __p0_162; \
+  int16x4_t __s1_162 = __p1_162; \
+  int32_t __ret_162; \
+  __ret_162 = vqdmullh_s16(__s0_162, vget_lane_s16(__s1_162, __p2_162)); \
+  __ret_162; \
+})
+#else
+#define vqdmullh_lane_s16(__p0_163, __p1_163, __p2_163) __extension__ ({ \
+  int16_t __s0_163 = __p0_163; \
+  int16x4_t __s1_163 = __p1_163; \
+  int16x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
+  int32_t __ret_163; \
+  __ret_163 = vqdmullh_s16(__s0_163, __noswap_vget_lane_s16(__rev1_163, __p2_163)); \
+  __ret_163; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmulls_laneq_s32(__p0_164, __p1_164, __p2_164) __extension__ ({ \
+  int32_t __s0_164 = __p0_164; \
+  int32x4_t __s1_164 = __p1_164; \
+  int64_t __ret_164; \
+  __ret_164 = vqdmulls_s32(__s0_164, vgetq_lane_s32(__s1_164, __p2_164)); \
+  __ret_164; \
+})
+#else
+#define vqdmulls_laneq_s32(__p0_165, __p1_165, __p2_165) __extension__ ({ \
+  int32_t __s0_165 = __p0_165; \
+  int32x4_t __s1_165 = __p1_165; \
+  int32x4_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 3, 2, 1, 0); \
+  int64_t __ret_165; \
+  __ret_165 = vqdmulls_s32(__s0_165, __noswap_vgetq_lane_s32(__rev1_165, __p2_165)); \
+  __ret_165; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqdmullh_laneq_s16(__p0_166, __p1_166, __p2_166) __extension__ ({ \
+  int16_t __s0_166 = __p0_166; \
+  int16x8_t __s1_166 = __p1_166; \
+  int32_t __ret_166; \
+  __ret_166 = vqdmullh_s16(__s0_166, vgetq_lane_s16(__s1_166, __p2_166)); \
   __ret_166; \
 })
 #else
-#define vqdmulls_lane_s32(__p0_167, __p1_167, __p2_167) __extension__ ({ \
-  int32_t __s0_167 = __p0_167; \
-  int32x2_t __s1_167 = __p1_167; \
-  int32x2_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 1, 0); \
-  int64_t __ret_167; \
-  __ret_167 = __noswap_vqdmulls_s32(__s0_167, __noswap_vget_lane_s32(__rev1_167, __p2_167)); \
+#define vqdmullh_laneq_s16(__p0_167, __p1_167, __p2_167) __extension__ ({ \
+  int16_t __s0_167 = __p0_167; \
+  int16x8_t __s1_167 = __p1_167; \
+  int16x8_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32_t __ret_167; \
+  __ret_167 = vqdmullh_s16(__s0_167, __noswap_vgetq_lane_s16(__rev1_167, __p2_167)); \
   __ret_167; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqdmullh_lane_s16(__p0_168, __p1_168, __p2_168) __extension__ ({ \
-  int16_t __s0_168 = __p0_168; \
-  int16x4_t __s1_168 = __p1_168; \
-  int32_t __ret_168; \
-  __ret_168 = vqdmullh_s16(__s0_168, vget_lane_s16(__s1_168, __p2_168)); \
-  __ret_168; \
-})
-#else
-#define vqdmullh_lane_s16(__p0_169, __p1_169, __p2_169) __extension__ ({ \
-  int16_t __s0_169 = __p0_169; \
-  int16x4_t __s1_169 = __p1_169; \
-  int16x4_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 3, 2, 1, 0); \
-  int32_t __ret_169; \
-  __ret_169 = __noswap_vqdmullh_s16(__s0_169, __noswap_vget_lane_s16(__rev1_169, __p2_169)); \
-  __ret_169; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulls_laneq_s32(__p0_170, __p1_170, __p2_170) __extension__ ({ \
-  int32_t __s0_170 = __p0_170; \
-  int32x4_t __s1_170 = __p1_170; \
-  int64_t __ret_170; \
-  __ret_170 = vqdmulls_s32(__s0_170, vgetq_lane_s32(__s1_170, __p2_170)); \
-  __ret_170; \
-})
-#else
-#define vqdmulls_laneq_s32(__p0_171, __p1_171, __p2_171) __extension__ ({ \
-  int32_t __s0_171 = __p0_171; \
-  int32x4_t __s1_171 = __p1_171; \
-  int32x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
-  int64_t __ret_171; \
-  __ret_171 = __noswap_vqdmulls_s32(__s0_171, __noswap_vgetq_lane_s32(__rev1_171, __p2_171)); \
-  __ret_171; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmullh_laneq_s16(__p0_172, __p1_172, __p2_172) __extension__ ({ \
-  int16_t __s0_172 = __p0_172; \
-  int16x8_t __s1_172 = __p1_172; \
-  int32_t __ret_172; \
-  __ret_172 = vqdmullh_s16(__s0_172, vgetq_lane_s16(__s1_172, __p2_172)); \
-  __ret_172; \
-})
-#else
-#define vqdmullh_laneq_s16(__p0_173, __p1_173, __p2_173) __extension__ ({ \
-  int16_t __s0_173 = __p0_173; \
-  int16x8_t __s1_173 = __p1_173; \
-  int16x8_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32_t __ret_173; \
-  __ret_173 = __noswap_vqdmullh_s16(__s0_173, __noswap_vgetq_lane_s16(__rev1_173, __p2_173)); \
-  __ret_173; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x2_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -62276,90 +52982,36 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqmovns_s32(int32_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqmovns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqmovnd_s64(int64_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqmovnd_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqmovnh_s16(int16_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqmovnh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqmovns_u32(uint32_t __p0) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
   return __ret;
 }
-#else
-__ai uint16_t vqmovns_u32(uint32_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqmovnd_u64(uint64_t __p0) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
   return __ret;
 }
-#else
-__ai uint32_t vqmovnd_u64(uint64_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqmovnh_u16(uint16_t __p0) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
   return __ret;
 }
-#else
-__ai uint8_t vqmovnh_u16(uint16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
   uint16x8_t __ret;
@@ -62462,48 +53114,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqmovuns_s32(int32_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqmovuns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqmovund_s64(int64_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqmovund_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqmovunh_s16(int16_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqmovunh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
   uint16x8_t __ret;
@@ -62571,191 +53196,118 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vqneg_s64(int64x1_t __p0) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vqneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqnegb_s8(int8_t __p0) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
   return __ret;
 }
-#else
-__ai int8_t vqnegb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqnegs_s32(int32_t __p0) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
   return __ret;
 }
-#else
-__ai int32_t vqnegs_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqnegd_s64(int64_t __p0) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
   return __ret;
 }
-#else
-__ai int64_t vqnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqnegh_s16(int16_t __p0) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
   return __ret;
 }
-#else
-__ai int16_t vqnegh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_lane_s32(__p0_168, __p1_168, __p2_168) __extension__ ({ \
+  int32_t __s0_168 = __p0_168; \
+  int32x2_t __s1_168 = __p1_168; \
+  int32_t __ret_168; \
+  __ret_168 = vqrdmulhs_s32(__s0_168, vget_lane_s32(__s1_168, __p2_168)); \
+  __ret_168; \
+})
 #else
-__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqrdmulhs_lane_s32(__p0_169, __p1_169, __p2_169) __extension__ ({ \
+  int32_t __s0_169 = __p0_169; \
+  int32x2_t __s1_169 = __p1_169; \
+  int32x2_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 1, 0); \
+  int32_t __ret_169; \
+  __ret_169 = vqrdmulhs_s32(__s0_169, __noswap_vget_lane_s32(__rev1_169, __p2_169)); \
+  __ret_169; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_lane_s32(__p0_174, __p1_174, __p2_174) __extension__ ({ \
-  int32_t __s0_174 = __p0_174; \
-  int32x2_t __s1_174 = __p1_174; \
-  int32_t __ret_174; \
-  __ret_174 = vqrdmulhs_s32(__s0_174, vget_lane_s32(__s1_174, __p2_174)); \
+#define vqrdmulhh_lane_s16(__p0_170, __p1_170, __p2_170) __extension__ ({ \
+  int16_t __s0_170 = __p0_170; \
+  int16x4_t __s1_170 = __p1_170; \
+  int16_t __ret_170; \
+  __ret_170 = vqrdmulhh_s16(__s0_170, vget_lane_s16(__s1_170, __p2_170)); \
+  __ret_170; \
+})
+#else
+#define vqrdmulhh_lane_s16(__p0_171, __p1_171, __p2_171) __extension__ ({ \
+  int16_t __s0_171 = __p0_171; \
+  int16x4_t __s1_171 = __p1_171; \
+  int16x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
+  int16_t __ret_171; \
+  __ret_171 = vqrdmulhh_s16(__s0_171, __noswap_vget_lane_s16(__rev1_171, __p2_171)); \
+  __ret_171; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhs_laneq_s32(__p0_172, __p1_172, __p2_172) __extension__ ({ \
+  int32_t __s0_172 = __p0_172; \
+  int32x4_t __s1_172 = __p1_172; \
+  int32_t __ret_172; \
+  __ret_172 = vqrdmulhs_s32(__s0_172, vgetq_lane_s32(__s1_172, __p2_172)); \
+  __ret_172; \
+})
+#else
+#define vqrdmulhs_laneq_s32(__p0_173, __p1_173, __p2_173) __extension__ ({ \
+  int32_t __s0_173 = __p0_173; \
+  int32x4_t __s1_173 = __p1_173; \
+  int32x4_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 3, 2, 1, 0); \
+  int32_t __ret_173; \
+  __ret_173 = vqrdmulhs_s32(__s0_173, __noswap_vgetq_lane_s32(__rev1_173, __p2_173)); \
+  __ret_173; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmulhh_laneq_s16(__p0_174, __p1_174, __p2_174) __extension__ ({ \
+  int16_t __s0_174 = __p0_174; \
+  int16x8_t __s1_174 = __p1_174; \
+  int16_t __ret_174; \
+  __ret_174 = vqrdmulhh_s16(__s0_174, vgetq_lane_s16(__s1_174, __p2_174)); \
   __ret_174; \
 })
 #else
-#define vqrdmulhs_lane_s32(__p0_175, __p1_175, __p2_175) __extension__ ({ \
-  int32_t __s0_175 = __p0_175; \
-  int32x2_t __s1_175 = __p1_175; \
-  int32x2_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 1, 0); \
-  int32_t __ret_175; \
-  __ret_175 = __noswap_vqrdmulhs_s32(__s0_175, __noswap_vget_lane_s32(__rev1_175, __p2_175)); \
+#define vqrdmulhh_laneq_s16(__p0_175, __p1_175, __p2_175) __extension__ ({ \
+  int16_t __s0_175 = __p0_175; \
+  int16x8_t __s1_175 = __p1_175; \
+  int16x8_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_175; \
+  __ret_175 = vqrdmulhh_s16(__s0_175, __noswap_vgetq_lane_s16(__rev1_175, __p2_175)); \
   __ret_175; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_lane_s16(__p0_176, __p1_176, __p2_176) __extension__ ({ \
-  int16_t __s0_176 = __p0_176; \
-  int16x4_t __s1_176 = __p1_176; \
-  int16_t __ret_176; \
-  __ret_176 = vqrdmulhh_s16(__s0_176, vget_lane_s16(__s1_176, __p2_176)); \
-  __ret_176; \
-})
-#else
-#define vqrdmulhh_lane_s16(__p0_177, __p1_177, __p2_177) __extension__ ({ \
-  int16_t __s0_177 = __p0_177; \
-  int16x4_t __s1_177 = __p1_177; \
-  int16x4_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 3, 2, 1, 0); \
-  int16_t __ret_177; \
-  __ret_177 = __noswap_vqrdmulhh_s16(__s0_177, __noswap_vget_lane_s16(__rev1_177, __p2_177)); \
-  __ret_177; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_laneq_s32(__p0_178, __p1_178, __p2_178) __extension__ ({ \
-  int32_t __s0_178 = __p0_178; \
-  int32x4_t __s1_178 = __p1_178; \
-  int32_t __ret_178; \
-  __ret_178 = vqrdmulhs_s32(__s0_178, vgetq_lane_s32(__s1_178, __p2_178)); \
-  __ret_178; \
-})
-#else
-#define vqrdmulhs_laneq_s32(__p0_179, __p1_179, __p2_179) __extension__ ({ \
-  int32_t __s0_179 = __p0_179; \
-  int32x4_t __s1_179 = __p1_179; \
-  int32x4_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 3, 2, 1, 0); \
-  int32_t __ret_179; \
-  __ret_179 = __noswap_vqrdmulhs_s32(__s0_179, __noswap_vgetq_lane_s32(__rev1_179, __p2_179)); \
-  __ret_179; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_laneq_s16(__p0_180, __p1_180, __p2_180) __extension__ ({ \
-  int16_t __s0_180 = __p0_180; \
-  int16x8_t __s1_180 = __p1_180; \
-  int16_t __ret_180; \
-  __ret_180 = vqrdmulhh_s16(__s0_180, vgetq_lane_s16(__s1_180, __p2_180)); \
-  __ret_180; \
-})
-#else
-#define vqrdmulhh_laneq_s16(__p0_181, __p1_181, __p2_181) __extension__ ({ \
-  int16_t __s0_181 = __p0_181; \
-  int16x8_t __s1_181 = __p1_181; \
-  int16x8_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_181; \
-  __ret_181 = __noswap_vqrdmulhh_s16(__s0_181, __noswap_vgetq_lane_s16(__rev1_181, __p2_181)); \
-  __ret_181; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
   int32x4_t __s0 = __p0; \
   int32x4_t __s1 = __p1; \
@@ -62839,1210 +53391,684 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u32(__p0_176, __p1_176, __p2_176) __extension__ ({ \
+  uint16x4_t __s0_176 = __p0_176; \
+  uint32x4_t __s1_176 = __p1_176; \
+  uint16x8_t __ret_176; \
+  __ret_176 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_176), (uint16x4_t)(vqrshrn_n_u32(__s1_176, __p2_176)))); \
+  __ret_176; \
+})
 #else
-__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
-  return __ret;
-}
+#define vqrshrn_high_n_u32(__p0_177, __p1_177, __p2_177) __extension__ ({ \
+  uint16x4_t __s0_177 = __p0_177; \
+  uint32x4_t __s1_177 = __p1_177; \
+  uint16x4_t __rev0_177;  __rev0_177 = __builtin_shufflevector(__s0_177, __s0_177, 3, 2, 1, 0); \
+  uint32x4_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 3, 2, 1, 0); \
+  uint16x8_t __ret_177; \
+  __ret_177 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_177), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_177, __p2_177)))); \
+  __ret_177 = __builtin_shufflevector(__ret_177, __ret_177, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_177; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u32(__p0_182, __p1_182, __p2_182) __extension__ ({ \
-  uint16x4_t __s0_182 = __p0_182; \
-  uint32x4_t __s1_182 = __p1_182; \
-  uint16x8_t __ret_182; \
-  __ret_182 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_182), (uint16x4_t)(vqrshrn_n_u32(__s1_182, __p2_182)))); \
+#define vqrshrn_high_n_u64(__p0_178, __p1_178, __p2_178) __extension__ ({ \
+  uint32x2_t __s0_178 = __p0_178; \
+  uint64x2_t __s1_178 = __p1_178; \
+  uint32x4_t __ret_178; \
+  __ret_178 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_178), (uint32x2_t)(vqrshrn_n_u64(__s1_178, __p2_178)))); \
+  __ret_178; \
+})
+#else
+#define vqrshrn_high_n_u64(__p0_179, __p1_179, __p2_179) __extension__ ({ \
+  uint32x2_t __s0_179 = __p0_179; \
+  uint64x2_t __s1_179 = __p1_179; \
+  uint32x2_t __rev0_179;  __rev0_179 = __builtin_shufflevector(__s0_179, __s0_179, 1, 0); \
+  uint64x2_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 1, 0); \
+  uint32x4_t __ret_179; \
+  __ret_179 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_179), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_179, __p2_179)))); \
+  __ret_179 = __builtin_shufflevector(__ret_179, __ret_179, 3, 2, 1, 0); \
+  __ret_179; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_u16(__p0_180, __p1_180, __p2_180) __extension__ ({ \
+  uint8x8_t __s0_180 = __p0_180; \
+  uint16x8_t __s1_180 = __p1_180; \
+  uint8x16_t __ret_180; \
+  __ret_180 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_180), (uint8x8_t)(vqrshrn_n_u16(__s1_180, __p2_180)))); \
+  __ret_180; \
+})
+#else
+#define vqrshrn_high_n_u16(__p0_181, __p1_181, __p2_181) __extension__ ({ \
+  uint8x8_t __s0_181 = __p0_181; \
+  uint16x8_t __s1_181 = __p1_181; \
+  uint8x8_t __rev0_181;  __rev0_181 = __builtin_shufflevector(__s0_181, __s0_181, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_181; \
+  __ret_181 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_181), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_181, __p2_181)))); \
+  __ret_181 = __builtin_shufflevector(__ret_181, __ret_181, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_181; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrshrn_high_n_s32(__p0_182, __p1_182, __p2_182) __extension__ ({ \
+  int16x4_t __s0_182 = __p0_182; \
+  int32x4_t __s1_182 = __p1_182; \
+  int16x8_t __ret_182; \
+  __ret_182 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_182), (int16x4_t)(vqrshrn_n_s32(__s1_182, __p2_182)))); \
   __ret_182; \
 })
 #else
-#define vqrshrn_high_n_u32(__p0_183, __p1_183, __p2_183) __extension__ ({ \
-  uint16x4_t __s0_183 = __p0_183; \
-  uint32x4_t __s1_183 = __p1_183; \
-  uint16x4_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 3, 2, 1, 0); \
-  uint32x4_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 3, 2, 1, 0); \
-  uint16x8_t __ret_183; \
-  __ret_183 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_183), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_183, __p2_183)))); \
+#define vqrshrn_high_n_s32(__p0_183, __p1_183, __p2_183) __extension__ ({ \
+  int16x4_t __s0_183 = __p0_183; \
+  int32x4_t __s1_183 = __p1_183; \
+  int16x4_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 3, 2, 1, 0); \
+  int32x4_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 3, 2, 1, 0); \
+  int16x8_t __ret_183; \
+  __ret_183 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_183), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_183, __p2_183)))); \
   __ret_183 = __builtin_shufflevector(__ret_183, __ret_183, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_183; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u64(__p0_184, __p1_184, __p2_184) __extension__ ({ \
-  uint32x2_t __s0_184 = __p0_184; \
-  uint64x2_t __s1_184 = __p1_184; \
-  uint32x4_t __ret_184; \
-  __ret_184 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_184), (uint32x2_t)(vqrshrn_n_u64(__s1_184, __p2_184)))); \
+#define vqrshrn_high_n_s64(__p0_184, __p1_184, __p2_184) __extension__ ({ \
+  int32x2_t __s0_184 = __p0_184; \
+  int64x2_t __s1_184 = __p1_184; \
+  int32x4_t __ret_184; \
+  __ret_184 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_184), (int32x2_t)(vqrshrn_n_s64(__s1_184, __p2_184)))); \
   __ret_184; \
 })
 #else
-#define vqrshrn_high_n_u64(__p0_185, __p1_185, __p2_185) __extension__ ({ \
-  uint32x2_t __s0_185 = __p0_185; \
-  uint64x2_t __s1_185 = __p1_185; \
-  uint32x2_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 1, 0); \
-  uint64x2_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 1, 0); \
-  uint32x4_t __ret_185; \
-  __ret_185 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_185), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_185, __p2_185)))); \
+#define vqrshrn_high_n_s64(__p0_185, __p1_185, __p2_185) __extension__ ({ \
+  int32x2_t __s0_185 = __p0_185; \
+  int64x2_t __s1_185 = __p1_185; \
+  int32x2_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 1, 0); \
+  int64x2_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 1, 0); \
+  int32x4_t __ret_185; \
+  __ret_185 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_185), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_185, __p2_185)))); \
   __ret_185 = __builtin_shufflevector(__ret_185, __ret_185, 3, 2, 1, 0); \
   __ret_185; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u16(__p0_186, __p1_186, __p2_186) __extension__ ({ \
-  uint8x8_t __s0_186 = __p0_186; \
-  uint16x8_t __s1_186 = __p1_186; \
-  uint8x16_t __ret_186; \
-  __ret_186 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_186), (uint8x8_t)(vqrshrn_n_u16(__s1_186, __p2_186)))); \
+#define vqrshrn_high_n_s16(__p0_186, __p1_186, __p2_186) __extension__ ({ \
+  int8x8_t __s0_186 = __p0_186; \
+  int16x8_t __s1_186 = __p1_186; \
+  int8x16_t __ret_186; \
+  __ret_186 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_186), (int8x8_t)(vqrshrn_n_s16(__s1_186, __p2_186)))); \
   __ret_186; \
 })
 #else
-#define vqrshrn_high_n_u16(__p0_187, __p1_187, __p2_187) __extension__ ({ \
-  uint8x8_t __s0_187 = __p0_187; \
-  uint16x8_t __s1_187 = __p1_187; \
-  uint8x8_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_187; \
-  __ret_187 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_187), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_187, __p2_187)))); \
+#define vqrshrn_high_n_s16(__p0_187, __p1_187, __p2_187) __extension__ ({ \
+  int8x8_t __s0_187 = __p0_187; \
+  int16x8_t __s1_187 = __p1_187; \
+  int8x8_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_187; \
+  __ret_187 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_187), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_187, __p2_187)))); \
   __ret_187 = __builtin_shufflevector(__ret_187, __ret_187, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_187; \
 })
 #endif
 
+#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s32(__p0_188, __p1_188, __p2_188) __extension__ ({ \
+#define vqrshrun_high_n_s32(__p0_188, __p1_188, __p2_188) __extension__ ({ \
   int16x4_t __s0_188 = __p0_188; \
   int32x4_t __s1_188 = __p1_188; \
   int16x8_t __ret_188; \
-  __ret_188 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_188), (int16x4_t)(vqrshrn_n_s32(__s1_188, __p2_188)))); \
+  __ret_188 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_188), (int16x4_t)(vqrshrun_n_s32(__s1_188, __p2_188)))); \
   __ret_188; \
 })
 #else
-#define vqrshrn_high_n_s32(__p0_189, __p1_189, __p2_189) __extension__ ({ \
+#define vqrshrun_high_n_s32(__p0_189, __p1_189, __p2_189) __extension__ ({ \
   int16x4_t __s0_189 = __p0_189; \
   int32x4_t __s1_189 = __p1_189; \
   int16x4_t __rev0_189;  __rev0_189 = __builtin_shufflevector(__s0_189, __s0_189, 3, 2, 1, 0); \
   int32x4_t __rev1_189;  __rev1_189 = __builtin_shufflevector(__s1_189, __s1_189, 3, 2, 1, 0); \
   int16x8_t __ret_189; \
-  __ret_189 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_189), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_189, __p2_189)))); \
+  __ret_189 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_189), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_189, __p2_189)))); \
   __ret_189 = __builtin_shufflevector(__ret_189, __ret_189, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_189; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s64(__p0_190, __p1_190, __p2_190) __extension__ ({ \
+#define vqrshrun_high_n_s64(__p0_190, __p1_190, __p2_190) __extension__ ({ \
   int32x2_t __s0_190 = __p0_190; \
   int64x2_t __s1_190 = __p1_190; \
   int32x4_t __ret_190; \
-  __ret_190 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_190), (int32x2_t)(vqrshrn_n_s64(__s1_190, __p2_190)))); \
+  __ret_190 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_190), (int32x2_t)(vqrshrun_n_s64(__s1_190, __p2_190)))); \
   __ret_190; \
 })
 #else
-#define vqrshrn_high_n_s64(__p0_191, __p1_191, __p2_191) __extension__ ({ \
+#define vqrshrun_high_n_s64(__p0_191, __p1_191, __p2_191) __extension__ ({ \
   int32x2_t __s0_191 = __p0_191; \
   int64x2_t __s1_191 = __p1_191; \
   int32x2_t __rev0_191;  __rev0_191 = __builtin_shufflevector(__s0_191, __s0_191, 1, 0); \
   int64x2_t __rev1_191;  __rev1_191 = __builtin_shufflevector(__s1_191, __s1_191, 1, 0); \
   int32x4_t __ret_191; \
-  __ret_191 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_191), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_191, __p2_191)))); \
+  __ret_191 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_191), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_191, __p2_191)))); \
   __ret_191 = __builtin_shufflevector(__ret_191, __ret_191, 3, 2, 1, 0); \
   __ret_191; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s16(__p0_192, __p1_192, __p2_192) __extension__ ({ \
+#define vqrshrun_high_n_s16(__p0_192, __p1_192, __p2_192) __extension__ ({ \
   int8x8_t __s0_192 = __p0_192; \
   int16x8_t __s1_192 = __p1_192; \
   int8x16_t __ret_192; \
-  __ret_192 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_192), (int8x8_t)(vqrshrn_n_s16(__s1_192, __p2_192)))); \
+  __ret_192 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_192), (int8x8_t)(vqrshrun_n_s16(__s1_192, __p2_192)))); \
   __ret_192; \
 })
 #else
-#define vqrshrn_high_n_s16(__p0_193, __p1_193, __p2_193) __extension__ ({ \
+#define vqrshrun_high_n_s16(__p0_193, __p1_193, __p2_193) __extension__ ({ \
   int8x8_t __s0_193 = __p0_193; \
   int16x8_t __s1_193 = __p1_193; \
   int8x8_t __rev0_193;  __rev0_193 = __builtin_shufflevector(__s0_193, __s0_193, 7, 6, 5, 4, 3, 2, 1, 0); \
   int16x8_t __rev1_193;  __rev1_193 = __builtin_shufflevector(__s1_193, __s1_193, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x16_t __ret_193; \
-  __ret_193 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_193), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_193, __p2_193)))); \
+  __ret_193 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_193), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_193, __p2_193)))); \
   __ret_193 = __builtin_shufflevector(__ret_193, __ret_193, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_193; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
+#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
+  __ret; \
+})
+__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
+  uint8_t __ret;
+  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
+  return __ret;
+}
+__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
+  uint32_t __ret;
+  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
+  return __ret;
+}
+__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
+  uint64_t __ret;
+  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
+  return __ret;
+}
+__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
+  uint16_t __ret;
+  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
+  return __ret;
+}
+__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
+  int8_t __ret;
+  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
+  return __ret;
+}
+__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
+  int32_t __ret;
+  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
+  return __ret;
+}
+__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
+  int64_t __ret;
+  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
+  return __ret;
+}
+__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
+  int16_t __ret;
+  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
+  return __ret;
+}
+#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
+  uint8_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
+  __ret; \
+})
+#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
+  __ret; \
+})
+#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
   int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
+  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
   __ret; \
 })
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
+#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
+  int8_t __s0 = __p0; \
   int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
+#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
   int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
   __ret; \
 })
-#endif
-
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s32(__p0_194, __p1_194, __p2_194) __extension__ ({ \
-  int16x4_t __s0_194 = __p0_194; \
-  int32x4_t __s1_194 = __p1_194; \
-  int16x8_t __ret_194; \
-  __ret_194 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_194), (int16x4_t)(vqrshrun_n_s32(__s1_194, __p2_194)))); \
+#define vqshrn_high_n_u32(__p0_194, __p1_194, __p2_194) __extension__ ({ \
+  uint16x4_t __s0_194 = __p0_194; \
+  uint32x4_t __s1_194 = __p1_194; \
+  uint16x8_t __ret_194; \
+  __ret_194 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_194), (uint16x4_t)(vqshrn_n_u32(__s1_194, __p2_194)))); \
   __ret_194; \
 })
 #else
-#define vqrshrun_high_n_s32(__p0_195, __p1_195, __p2_195) __extension__ ({ \
-  int16x4_t __s0_195 = __p0_195; \
-  int32x4_t __s1_195 = __p1_195; \
-  int16x4_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 3, 2, 1, 0); \
-  int32x4_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 3, 2, 1, 0); \
-  int16x8_t __ret_195; \
-  __ret_195 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_195), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_195, __p2_195)))); \
+#define vqshrn_high_n_u32(__p0_195, __p1_195, __p2_195) __extension__ ({ \
+  uint16x4_t __s0_195 = __p0_195; \
+  uint32x4_t __s1_195 = __p1_195; \
+  uint16x4_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 3, 2, 1, 0); \
+  uint32x4_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 3, 2, 1, 0); \
+  uint16x8_t __ret_195; \
+  __ret_195 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_195), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_195, __p2_195)))); \
   __ret_195 = __builtin_shufflevector(__ret_195, __ret_195, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_195; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s64(__p0_196, __p1_196, __p2_196) __extension__ ({ \
-  int32x2_t __s0_196 = __p0_196; \
-  int64x2_t __s1_196 = __p1_196; \
-  int32x4_t __ret_196; \
-  __ret_196 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_196), (int32x2_t)(vqrshrun_n_s64(__s1_196, __p2_196)))); \
+#define vqshrn_high_n_u64(__p0_196, __p1_196, __p2_196) __extension__ ({ \
+  uint32x2_t __s0_196 = __p0_196; \
+  uint64x2_t __s1_196 = __p1_196; \
+  uint32x4_t __ret_196; \
+  __ret_196 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_196), (uint32x2_t)(vqshrn_n_u64(__s1_196, __p2_196)))); \
   __ret_196; \
 })
 #else
-#define vqrshrun_high_n_s64(__p0_197, __p1_197, __p2_197) __extension__ ({ \
-  int32x2_t __s0_197 = __p0_197; \
-  int64x2_t __s1_197 = __p1_197; \
-  int32x2_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 1, 0); \
-  int64x2_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 1, 0); \
-  int32x4_t __ret_197; \
-  __ret_197 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_197), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_197, __p2_197)))); \
+#define vqshrn_high_n_u64(__p0_197, __p1_197, __p2_197) __extension__ ({ \
+  uint32x2_t __s0_197 = __p0_197; \
+  uint64x2_t __s1_197 = __p1_197; \
+  uint32x2_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 1, 0); \
+  uint64x2_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 1, 0); \
+  uint32x4_t __ret_197; \
+  __ret_197 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_197), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_197, __p2_197)))); \
   __ret_197 = __builtin_shufflevector(__ret_197, __ret_197, 3, 2, 1, 0); \
   __ret_197; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s16(__p0_198, __p1_198, __p2_198) __extension__ ({ \
-  int8x8_t __s0_198 = __p0_198; \
-  int16x8_t __s1_198 = __p1_198; \
-  int8x16_t __ret_198; \
-  __ret_198 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_198), (int8x8_t)(vqrshrun_n_s16(__s1_198, __p2_198)))); \
+#define vqshrn_high_n_u16(__p0_198, __p1_198, __p2_198) __extension__ ({ \
+  uint8x8_t __s0_198 = __p0_198; \
+  uint16x8_t __s1_198 = __p1_198; \
+  uint8x16_t __ret_198; \
+  __ret_198 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_198), (uint8x8_t)(vqshrn_n_u16(__s1_198, __p2_198)))); \
   __ret_198; \
 })
 #else
-#define vqrshrun_high_n_s16(__p0_199, __p1_199, __p2_199) __extension__ ({ \
-  int8x8_t __s0_199 = __p0_199; \
-  int16x8_t __s1_199 = __p1_199; \
-  int8x8_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_199; \
-  __ret_199 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_199), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_199, __p2_199)))); \
+#define vqshrn_high_n_u16(__p0_199, __p1_199, __p2_199) __extension__ ({ \
+  uint8x8_t __s0_199 = __p0_199; \
+  uint16x8_t __s1_199 = __p1_199; \
+  uint8x8_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_199; \
+  __ret_199 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_199), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_199, __p2_199)))); \
   __ret_199 = __builtin_shufflevector(__ret_199, __ret_199, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_199; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u32(__p0_200, __p1_200, __p2_200) __extension__ ({ \
-  uint16x4_t __s0_200 = __p0_200; \
-  uint32x4_t __s1_200 = __p1_200; \
-  uint16x8_t __ret_200; \
-  __ret_200 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_200), (uint16x4_t)(vqshrn_n_u32(__s1_200, __p2_200)))); \
+#define vqshrn_high_n_s32(__p0_200, __p1_200, __p2_200) __extension__ ({ \
+  int16x4_t __s0_200 = __p0_200; \
+  int32x4_t __s1_200 = __p1_200; \
+  int16x8_t __ret_200; \
+  __ret_200 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_200), (int16x4_t)(vqshrn_n_s32(__s1_200, __p2_200)))); \
   __ret_200; \
 })
 #else
-#define vqshrn_high_n_u32(__p0_201, __p1_201, __p2_201) __extension__ ({ \
-  uint16x4_t __s0_201 = __p0_201; \
-  uint32x4_t __s1_201 = __p1_201; \
-  uint16x4_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 3, 2, 1, 0); \
-  uint32x4_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 3, 2, 1, 0); \
-  uint16x8_t __ret_201; \
-  __ret_201 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_201), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_201, __p2_201)))); \
+#define vqshrn_high_n_s32(__p0_201, __p1_201, __p2_201) __extension__ ({ \
+  int16x4_t __s0_201 = __p0_201; \
+  int32x4_t __s1_201 = __p1_201; \
+  int16x4_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 3, 2, 1, 0); \
+  int32x4_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 3, 2, 1, 0); \
+  int16x8_t __ret_201; \
+  __ret_201 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_201), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_201, __p2_201)))); \
   __ret_201 = __builtin_shufflevector(__ret_201, __ret_201, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_201; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u64(__p0_202, __p1_202, __p2_202) __extension__ ({ \
-  uint32x2_t __s0_202 = __p0_202; \
-  uint64x2_t __s1_202 = __p1_202; \
-  uint32x4_t __ret_202; \
-  __ret_202 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_202), (uint32x2_t)(vqshrn_n_u64(__s1_202, __p2_202)))); \
+#define vqshrn_high_n_s64(__p0_202, __p1_202, __p2_202) __extension__ ({ \
+  int32x2_t __s0_202 = __p0_202; \
+  int64x2_t __s1_202 = __p1_202; \
+  int32x4_t __ret_202; \
+  __ret_202 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_202), (int32x2_t)(vqshrn_n_s64(__s1_202, __p2_202)))); \
   __ret_202; \
 })
 #else
-#define vqshrn_high_n_u64(__p0_203, __p1_203, __p2_203) __extension__ ({ \
-  uint32x2_t __s0_203 = __p0_203; \
-  uint64x2_t __s1_203 = __p1_203; \
-  uint32x2_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 1, 0); \
-  uint64x2_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 1, 0); \
-  uint32x4_t __ret_203; \
-  __ret_203 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_203), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_203, __p2_203)))); \
+#define vqshrn_high_n_s64(__p0_203, __p1_203, __p2_203) __extension__ ({ \
+  int32x2_t __s0_203 = __p0_203; \
+  int64x2_t __s1_203 = __p1_203; \
+  int32x2_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 1, 0); \
+  int64x2_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 1, 0); \
+  int32x4_t __ret_203; \
+  __ret_203 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_203), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_203, __p2_203)))); \
   __ret_203 = __builtin_shufflevector(__ret_203, __ret_203, 3, 2, 1, 0); \
   __ret_203; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u16(__p0_204, __p1_204, __p2_204) __extension__ ({ \
-  uint8x8_t __s0_204 = __p0_204; \
-  uint16x8_t __s1_204 = __p1_204; \
-  uint8x16_t __ret_204; \
-  __ret_204 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_204), (uint8x8_t)(vqshrn_n_u16(__s1_204, __p2_204)))); \
+#define vqshrn_high_n_s16(__p0_204, __p1_204, __p2_204) __extension__ ({ \
+  int8x8_t __s0_204 = __p0_204; \
+  int16x8_t __s1_204 = __p1_204; \
+  int8x16_t __ret_204; \
+  __ret_204 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_204), (int8x8_t)(vqshrn_n_s16(__s1_204, __p2_204)))); \
   __ret_204; \
 })
 #else
-#define vqshrn_high_n_u16(__p0_205, __p1_205, __p2_205) __extension__ ({ \
-  uint8x8_t __s0_205 = __p0_205; \
-  uint16x8_t __s1_205 = __p1_205; \
-  uint8x8_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_205; \
-  __ret_205 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_205), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_205, __p2_205)))); \
+#define vqshrn_high_n_s16(__p0_205, __p1_205, __p2_205) __extension__ ({ \
+  int8x8_t __s0_205 = __p0_205; \
+  int16x8_t __s1_205 = __p1_205; \
+  int8x8_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_205; \
+  __ret_205 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_205), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_205, __p2_205)))); \
   __ret_205 = __builtin_shufflevector(__ret_205, __ret_205, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_205; \
 })
 #endif
 
+#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
+  uint32_t __s0 = __p0; \
+  uint16_t __ret; \
+  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint32_t __ret; \
+  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
+  uint16_t __s0 = __p0; \
+  uint8_t __ret; \
+  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
+  __ret; \
+})
+#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
+  int32_t __s0 = __p0; \
+  int16_t __ret; \
+  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int32_t __ret; \
+  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
+  __ret; \
+})
+#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
+  int16_t __s0 = __p0; \
+  int8_t __ret; \
+  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s32(__p0_206, __p1_206, __p2_206) __extension__ ({ \
+#define vqshrun_high_n_s32(__p0_206, __p1_206, __p2_206) __extension__ ({ \
   int16x4_t __s0_206 = __p0_206; \
   int32x4_t __s1_206 = __p1_206; \
   int16x8_t __ret_206; \
-  __ret_206 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_206), (int16x4_t)(vqshrn_n_s32(__s1_206, __p2_206)))); \
+  __ret_206 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_206), (int16x4_t)(vqshrun_n_s32(__s1_206, __p2_206)))); \
   __ret_206; \
 })
 #else
-#define vqshrn_high_n_s32(__p0_207, __p1_207, __p2_207) __extension__ ({ \
+#define vqshrun_high_n_s32(__p0_207, __p1_207, __p2_207) __extension__ ({ \
   int16x4_t __s0_207 = __p0_207; \
   int32x4_t __s1_207 = __p1_207; \
   int16x4_t __rev0_207;  __rev0_207 = __builtin_shufflevector(__s0_207, __s0_207, 3, 2, 1, 0); \
   int32x4_t __rev1_207;  __rev1_207 = __builtin_shufflevector(__s1_207, __s1_207, 3, 2, 1, 0); \
   int16x8_t __ret_207; \
-  __ret_207 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_207), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_207, __p2_207)))); \
+  __ret_207 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_207), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_207, __p2_207)))); \
   __ret_207 = __builtin_shufflevector(__ret_207, __ret_207, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_207; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s64(__p0_208, __p1_208, __p2_208) __extension__ ({ \
+#define vqshrun_high_n_s64(__p0_208, __p1_208, __p2_208) __extension__ ({ \
   int32x2_t __s0_208 = __p0_208; \
   int64x2_t __s1_208 = __p1_208; \
   int32x4_t __ret_208; \
-  __ret_208 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_208), (int32x2_t)(vqshrn_n_s64(__s1_208, __p2_208)))); \
+  __ret_208 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_208), (int32x2_t)(vqshrun_n_s64(__s1_208, __p2_208)))); \
   __ret_208; \
 })
 #else
-#define vqshrn_high_n_s64(__p0_209, __p1_209, __p2_209) __extension__ ({ \
+#define vqshrun_high_n_s64(__p0_209, __p1_209, __p2_209) __extension__ ({ \
   int32x2_t __s0_209 = __p0_209; \
   int64x2_t __s1_209 = __p1_209; \
   int32x2_t __rev0_209;  __rev0_209 = __builtin_shufflevector(__s0_209, __s0_209, 1, 0); \
   int64x2_t __rev1_209;  __rev1_209 = __builtin_shufflevector(__s1_209, __s1_209, 1, 0); \
   int32x4_t __ret_209; \
-  __ret_209 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_209), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_209, __p2_209)))); \
+  __ret_209 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_209), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_209, __p2_209)))); \
   __ret_209 = __builtin_shufflevector(__ret_209, __ret_209, 3, 2, 1, 0); \
   __ret_209; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s16(__p0_210, __p1_210, __p2_210) __extension__ ({ \
+#define vqshrun_high_n_s16(__p0_210, __p1_210, __p2_210) __extension__ ({ \
   int8x8_t __s0_210 = __p0_210; \
   int16x8_t __s1_210 = __p1_210; \
   int8x16_t __ret_210; \
-  __ret_210 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_210), (int8x8_t)(vqshrn_n_s16(__s1_210, __p2_210)))); \
+  __ret_210 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_210), (int8x8_t)(vqshrun_n_s16(__s1_210, __p2_210)))); \
   __ret_210; \
 })
 #else
-#define vqshrn_high_n_s16(__p0_211, __p1_211, __p2_211) __extension__ ({ \
+#define vqshrun_high_n_s16(__p0_211, __p1_211, __p2_211) __extension__ ({ \
   int8x8_t __s0_211 = __p0_211; \
   int16x8_t __s1_211 = __p1_211; \
   int8x8_t __rev0_211;  __rev0_211 = __builtin_shufflevector(__s0_211, __s0_211, 7, 6, 5, 4, 3, 2, 1, 0); \
   int16x8_t __rev1_211;  __rev1_211 = __builtin_shufflevector(__s1_211, __s1_211, 7, 6, 5, 4, 3, 2, 1, 0); \
   int8x16_t __ret_211; \
-  __ret_211 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_211), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_211, __p2_211)))); \
+  __ret_211 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_211), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_211, __p2_211)))); \
   __ret_211 = __builtin_shufflevector(__ret_211, __ret_211, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_211; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s32(__p0_212, __p1_212, __p2_212) __extension__ ({ \
-  int16x4_t __s0_212 = __p0_212; \
-  int32x4_t __s1_212 = __p1_212; \
-  int16x8_t __ret_212; \
-  __ret_212 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_212), (int16x4_t)(vqshrun_n_s32(__s1_212, __p2_212)))); \
-  __ret_212; \
-})
-#else
-#define vqshrun_high_n_s32(__p0_213, __p1_213, __p2_213) __extension__ ({ \
-  int16x4_t __s0_213 = __p0_213; \
-  int32x4_t __s1_213 = __p1_213; \
-  int16x4_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 3, 2, 1, 0); \
-  int32x4_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 3, 2, 1, 0); \
-  int16x8_t __ret_213; \
-  __ret_213 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_213), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_213, __p2_213)))); \
-  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_213; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s64(__p0_214, __p1_214, __p2_214) __extension__ ({ \
-  int32x2_t __s0_214 = __p0_214; \
-  int64x2_t __s1_214 = __p1_214; \
-  int32x4_t __ret_214; \
-  __ret_214 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_214), (int32x2_t)(vqshrun_n_s64(__s1_214, __p2_214)))); \
-  __ret_214; \
-})
-#else
-#define vqshrun_high_n_s64(__p0_215, __p1_215, __p2_215) __extension__ ({ \
-  int32x2_t __s0_215 = __p0_215; \
-  int64x2_t __s1_215 = __p1_215; \
-  int32x2_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 1, 0); \
-  int64x2_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 1, 0); \
-  int32x4_t __ret_215; \
-  __ret_215 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_215), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_215, __p2_215)))); \
-  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 3, 2, 1, 0); \
-  __ret_215; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s16(__p0_216, __p1_216, __p2_216) __extension__ ({ \
-  int8x8_t __s0_216 = __p0_216; \
-  int16x8_t __s1_216 = __p1_216; \
-  int8x16_t __ret_216; \
-  __ret_216 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_216), (int8x8_t)(vqshrun_n_s16(__s1_216, __p2_216)))); \
-  __ret_216; \
-})
-#else
-#define vqshrun_high_n_s16(__p0_217, __p1_217, __p2_217) __extension__ ({ \
-  int8x8_t __s0_217 = __p0_217; \
-  int16x8_t __s1_217 = __p1_217; \
-  int8x8_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_217; \
-  __ret_217 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_217), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_217, __p2_217)))); \
-  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_217; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
   int32_t __s0 = __p0; \
   int16_t __ret; \
   __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int32_t __ret; \
   __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
   int16_t __s0 = __p0; \
   int8_t __ret; \
   __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
   __ret; \
 })
-#else
-#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
   poly8x8_t __ret;
@@ -65211,48 +55237,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrecpe_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecped_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrecped_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpes_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrecpes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -65270,263 +55269,180 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrecpxd_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrecpxd_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrecpxs_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrecpxs_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
   __ret; \
 })
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u32(__p0_212, __p1_212, __p2_212) __extension__ ({ \
+  uint16x4_t __s0_212 = __p0_212; \
+  uint32x4_t __s1_212 = __p1_212; \
+  uint16x8_t __ret_212; \
+  __ret_212 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_212), (uint16x4_t)(vrshrn_n_u32(__s1_212, __p2_212)))); \
+  __ret_212; \
+})
 #else
-#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
-  __ret; \
+#define vrshrn_high_n_u32(__p0_213, __p1_213, __p2_213) __extension__ ({ \
+  uint16x4_t __s0_213 = __p0_213; \
+  uint32x4_t __s1_213 = __p1_213; \
+  uint16x4_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 3, 2, 1, 0); \
+  uint32x4_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 3, 2, 1, 0); \
+  uint16x8_t __ret_213; \
+  __ret_213 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_213), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_213, __p2_213)))); \
+  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_213; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u32(__p0_218, __p1_218, __p2_218) __extension__ ({ \
-  uint16x4_t __s0_218 = __p0_218; \
-  uint32x4_t __s1_218 = __p1_218; \
-  uint16x8_t __ret_218; \
-  __ret_218 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_218), (uint16x4_t)(vrshrn_n_u32(__s1_218, __p2_218)))); \
+#define vrshrn_high_n_u64(__p0_214, __p1_214, __p2_214) __extension__ ({ \
+  uint32x2_t __s0_214 = __p0_214; \
+  uint64x2_t __s1_214 = __p1_214; \
+  uint32x4_t __ret_214; \
+  __ret_214 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_214), (uint32x2_t)(vrshrn_n_u64(__s1_214, __p2_214)))); \
+  __ret_214; \
+})
+#else
+#define vrshrn_high_n_u64(__p0_215, __p1_215, __p2_215) __extension__ ({ \
+  uint32x2_t __s0_215 = __p0_215; \
+  uint64x2_t __s1_215 = __p1_215; \
+  uint32x2_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 1, 0); \
+  uint64x2_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 1, 0); \
+  uint32x4_t __ret_215; \
+  __ret_215 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_215), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_215, __p2_215)))); \
+  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 3, 2, 1, 0); \
+  __ret_215; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_u16(__p0_216, __p1_216, __p2_216) __extension__ ({ \
+  uint8x8_t __s0_216 = __p0_216; \
+  uint16x8_t __s1_216 = __p1_216; \
+  uint8x16_t __ret_216; \
+  __ret_216 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_216), (uint8x8_t)(vrshrn_n_u16(__s1_216, __p2_216)))); \
+  __ret_216; \
+})
+#else
+#define vrshrn_high_n_u16(__p0_217, __p1_217, __p2_217) __extension__ ({ \
+  uint8x8_t __s0_217 = __p0_217; \
+  uint16x8_t __s1_217 = __p1_217; \
+  uint8x8_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_217; \
+  __ret_217 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_217), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_217, __p2_217)))); \
+  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_217; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vrshrn_high_n_s32(__p0_218, __p1_218, __p2_218) __extension__ ({ \
+  int16x4_t __s0_218 = __p0_218; \
+  int32x4_t __s1_218 = __p1_218; \
+  int16x8_t __ret_218; \
+  __ret_218 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_218), (int16x4_t)(vrshrn_n_s32(__s1_218, __p2_218)))); \
   __ret_218; \
 })
 #else
-#define vrshrn_high_n_u32(__p0_219, __p1_219, __p2_219) __extension__ ({ \
-  uint16x4_t __s0_219 = __p0_219; \
-  uint32x4_t __s1_219 = __p1_219; \
-  uint16x4_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 3, 2, 1, 0); \
-  uint32x4_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 3, 2, 1, 0); \
-  uint16x8_t __ret_219; \
-  __ret_219 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_219), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_219, __p2_219)))); \
+#define vrshrn_high_n_s32(__p0_219, __p1_219, __p2_219) __extension__ ({ \
+  int16x4_t __s0_219 = __p0_219; \
+  int32x4_t __s1_219 = __p1_219; \
+  int16x4_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 3, 2, 1, 0); \
+  int32x4_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 3, 2, 1, 0); \
+  int16x8_t __ret_219; \
+  __ret_219 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_219), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_219, __p2_219)))); \
   __ret_219 = __builtin_shufflevector(__ret_219, __ret_219, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_219; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u64(__p0_220, __p1_220, __p2_220) __extension__ ({ \
-  uint32x2_t __s0_220 = __p0_220; \
-  uint64x2_t __s1_220 = __p1_220; \
-  uint32x4_t __ret_220; \
-  __ret_220 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_220), (uint32x2_t)(vrshrn_n_u64(__s1_220, __p2_220)))); \
+#define vrshrn_high_n_s64(__p0_220, __p1_220, __p2_220) __extension__ ({ \
+  int32x2_t __s0_220 = __p0_220; \
+  int64x2_t __s1_220 = __p1_220; \
+  int32x4_t __ret_220; \
+  __ret_220 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_220), (int32x2_t)(vrshrn_n_s64(__s1_220, __p2_220)))); \
   __ret_220; \
 })
 #else
-#define vrshrn_high_n_u64(__p0_221, __p1_221, __p2_221) __extension__ ({ \
-  uint32x2_t __s0_221 = __p0_221; \
-  uint64x2_t __s1_221 = __p1_221; \
-  uint32x2_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 1, 0); \
-  uint64x2_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 1, 0); \
-  uint32x4_t __ret_221; \
-  __ret_221 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_221), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_221, __p2_221)))); \
+#define vrshrn_high_n_s64(__p0_221, __p1_221, __p2_221) __extension__ ({ \
+  int32x2_t __s0_221 = __p0_221; \
+  int64x2_t __s1_221 = __p1_221; \
+  int32x2_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 1, 0); \
+  int64x2_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 1, 0); \
+  int32x4_t __ret_221; \
+  __ret_221 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_221), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_221, __p2_221)))); \
   __ret_221 = __builtin_shufflevector(__ret_221, __ret_221, 3, 2, 1, 0); \
   __ret_221; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
-  uint8x8_t __s0_222 = __p0_222; \
-  uint16x8_t __s1_222 = __p1_222; \
-  uint8x16_t __ret_222; \
-  __ret_222 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_222), (uint8x8_t)(vrshrn_n_u16(__s1_222, __p2_222)))); \
+#define vrshrn_high_n_s16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
+  int8x8_t __s0_222 = __p0_222; \
+  int16x8_t __s1_222 = __p1_222; \
+  int8x16_t __ret_222; \
+  __ret_222 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_222), (int8x8_t)(vrshrn_n_s16(__s1_222, __p2_222)))); \
   __ret_222; \
 })
 #else
-#define vrshrn_high_n_u16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
-  uint8x8_t __s0_223 = __p0_223; \
-  uint16x8_t __s1_223 = __p1_223; \
-  uint8x8_t __rev0_223;  __rev0_223 = __builtin_shufflevector(__s0_223, __s0_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_223; \
-  __ret_223 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_223), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_223, __p2_223)))); \
+#define vrshrn_high_n_s16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
+  int8x8_t __s0_223 = __p0_223; \
+  int16x8_t __s1_223 = __p1_223; \
+  int8x8_t __rev0_223;  __rev0_223 = __builtin_shufflevector(__s0_223, __s0_223, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_223; \
+  __ret_223 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_223), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_223, __p2_223)))); \
   __ret_223 = __builtin_shufflevector(__ret_223, __ret_223, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_223; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s32(__p0_224, __p1_224, __p2_224) __extension__ ({ \
-  int16x4_t __s0_224 = __p0_224; \
-  int32x4_t __s1_224 = __p1_224; \
-  int16x8_t __ret_224; \
-  __ret_224 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_224), (int16x4_t)(vrshrn_n_s32(__s1_224, __p2_224)))); \
-  __ret_224; \
-})
-#else
-#define vrshrn_high_n_s32(__p0_225, __p1_225, __p2_225) __extension__ ({ \
-  int16x4_t __s0_225 = __p0_225; \
-  int32x4_t __s1_225 = __p1_225; \
-  int16x4_t __rev0_225;  __rev0_225 = __builtin_shufflevector(__s0_225, __s0_225, 3, 2, 1, 0); \
-  int32x4_t __rev1_225;  __rev1_225 = __builtin_shufflevector(__s1_225, __s1_225, 3, 2, 1, 0); \
-  int16x8_t __ret_225; \
-  __ret_225 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_225), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_225, __p2_225)))); \
-  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_225; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s64(__p0_226, __p1_226, __p2_226) __extension__ ({ \
-  int32x2_t __s0_226 = __p0_226; \
-  int64x2_t __s1_226 = __p1_226; \
-  int32x4_t __ret_226; \
-  __ret_226 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_226), (int32x2_t)(vrshrn_n_s64(__s1_226, __p2_226)))); \
-  __ret_226; \
-})
-#else
-#define vrshrn_high_n_s64(__p0_227, __p1_227, __p2_227) __extension__ ({ \
-  int32x2_t __s0_227 = __p0_227; \
-  int64x2_t __s1_227 = __p1_227; \
-  int32x2_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 1, 0); \
-  int64x2_t __rev1_227;  __rev1_227 = __builtin_shufflevector(__s1_227, __s1_227, 1, 0); \
-  int32x4_t __ret_227; \
-  __ret_227 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_227), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_227, __p2_227)))); \
-  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 3, 2, 1, 0); \
-  __ret_227; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s16(__p0_228, __p1_228, __p2_228) __extension__ ({ \
-  int8x8_t __s0_228 = __p0_228; \
-  int16x8_t __s1_228 = __p1_228; \
-  int8x16_t __ret_228; \
-  __ret_228 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_228), (int8x8_t)(vrshrn_n_s16(__s1_228, __p2_228)))); \
-  __ret_228; \
-})
-#else
-#define vrshrn_high_n_s16(__p0_229, __p1_229, __p2_229) __extension__ ({ \
-  int8x8_t __s0_229 = __p0_229; \
-  int16x8_t __s1_229 = __p1_229; \
-  int8x8_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_229;  __rev1_229 = __builtin_shufflevector(__s1_229, __s1_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_229; \
-  __ret_229 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_229), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_229, __p2_229)))); \
-  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_229; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
   float64x2_t __ret;
   __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 42);
@@ -65542,48 +55458,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrsqrted_f64(float64_t __p0) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
   return __ret;
 }
-#else
-__ai float64_t vrsqrted_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrsqrtes_f32(float32_t __p0) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
   return __ret;
 }
-#else
-__ai float32_t vrsqrtes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -65601,49 +55490,21 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
   float64_t __ret;
   __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
   return __ret;
 }
-#else
-__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
   float32_t __ret;
   __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
   return __ret;
 }
-#else
-__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -65651,17 +55512,6 @@
   __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -65669,16 +55519,6 @@
   __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -65787,7 +55627,6 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -65795,23 +55634,6 @@
   __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64_t __s0 = __p0; \
@@ -65866,7 +55688,6 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64_t __s0 = __p0; \
   float64x1_t __s1 = __p1; \
@@ -65874,350 +55695,274 @@
   __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
   __ret; \
 })
-#else
-#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vshld_n_u64(__p0, __p1) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __ret; \
   __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
   __ret; \
 })
-#else
-#define vshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vshld_n_s64(__p0, __p1) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __ret; \
   __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
   __ret; \
 })
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u8(__p0_224, __p1_224) __extension__ ({ \
+  uint8x16_t __s0_224 = __p0_224; \
+  uint16x8_t __ret_224; \
+  __ret_224 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_224), __p1_224)); \
+  __ret_224; \
+})
 #else
-#define vshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
-  __ret; \
+#define vshll_high_n_u8(__p0_225, __p1_225) __extension__ ({ \
+  uint8x16_t __s0_225 = __p0_225; \
+  uint8x16_t __rev0_225;  __rev0_225 = __builtin_shufflevector(__s0_225, __s0_225, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __ret_225; \
+  __ret_225 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_225), __p1_225)); \
+  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_225; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u8(__p0_230, __p1_230) __extension__ ({ \
-  uint8x16_t __s0_230 = __p0_230; \
-  uint16x8_t __ret_230; \
-  __ret_230 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_230), __p1_230)); \
+#define vshll_high_n_u32(__p0_226, __p1_226) __extension__ ({ \
+  uint32x4_t __s0_226 = __p0_226; \
+  uint64x2_t __ret_226; \
+  __ret_226 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_226), __p1_226)); \
+  __ret_226; \
+})
+#else
+#define vshll_high_n_u32(__p0_227, __p1_227) __extension__ ({ \
+  uint32x4_t __s0_227 = __p0_227; \
+  uint32x4_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 3, 2, 1, 0); \
+  uint64x2_t __ret_227; \
+  __ret_227 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_227), __p1_227)); \
+  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 1, 0); \
+  __ret_227; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_u16(__p0_228, __p1_228) __extension__ ({ \
+  uint16x8_t __s0_228 = __p0_228; \
+  uint32x4_t __ret_228; \
+  __ret_228 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_228), __p1_228)); \
+  __ret_228; \
+})
+#else
+#define vshll_high_n_u16(__p0_229, __p1_229) __extension__ ({ \
+  uint16x8_t __s0_229 = __p0_229; \
+  uint16x8_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint32x4_t __ret_229; \
+  __ret_229 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_229), __p1_229)); \
+  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 3, 2, 1, 0); \
+  __ret_229; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vshll_high_n_s8(__p0_230, __p1_230) __extension__ ({ \
+  int8x16_t __s0_230 = __p0_230; \
+  int16x8_t __ret_230; \
+  __ret_230 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_230), __p1_230)); \
   __ret_230; \
 })
 #else
-#define vshll_high_n_u8(__p0_231, __p1_231) __extension__ ({ \
-  uint8x16_t __s0_231 = __p0_231; \
-  uint8x16_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_231; \
-  __ret_231 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_231), __p1_231)); \
+#define vshll_high_n_s8(__p0_231, __p1_231) __extension__ ({ \
+  int8x16_t __s0_231 = __p0_231; \
+  int8x16_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __ret_231; \
+  __ret_231 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_231), __p1_231)); \
   __ret_231 = __builtin_shufflevector(__ret_231, __ret_231, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_231; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u32(__p0_232, __p1_232) __extension__ ({ \
-  uint32x4_t __s0_232 = __p0_232; \
-  uint64x2_t __ret_232; \
-  __ret_232 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_232), __p1_232)); \
+#define vshll_high_n_s32(__p0_232, __p1_232) __extension__ ({ \
+  int32x4_t __s0_232 = __p0_232; \
+  int64x2_t __ret_232; \
+  __ret_232 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_232), __p1_232)); \
   __ret_232; \
 })
 #else
-#define vshll_high_n_u32(__p0_233, __p1_233) __extension__ ({ \
-  uint32x4_t __s0_233 = __p0_233; \
-  uint32x4_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 3, 2, 1, 0); \
-  uint64x2_t __ret_233; \
-  __ret_233 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_233), __p1_233)); \
+#define vshll_high_n_s32(__p0_233, __p1_233) __extension__ ({ \
+  int32x4_t __s0_233 = __p0_233; \
+  int32x4_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 3, 2, 1, 0); \
+  int64x2_t __ret_233; \
+  __ret_233 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_233), __p1_233)); \
   __ret_233 = __builtin_shufflevector(__ret_233, __ret_233, 1, 0); \
   __ret_233; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u16(__p0_234, __p1_234) __extension__ ({ \
-  uint16x8_t __s0_234 = __p0_234; \
-  uint32x4_t __ret_234; \
-  __ret_234 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_234), __p1_234)); \
+#define vshll_high_n_s16(__p0_234, __p1_234) __extension__ ({ \
+  int16x8_t __s0_234 = __p0_234; \
+  int32x4_t __ret_234; \
+  __ret_234 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_234), __p1_234)); \
   __ret_234; \
 })
 #else
-#define vshll_high_n_u16(__p0_235, __p1_235) __extension__ ({ \
-  uint16x8_t __s0_235 = __p0_235; \
-  uint16x8_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret_235; \
-  __ret_235 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_235), __p1_235)); \
+#define vshll_high_n_s16(__p0_235, __p1_235) __extension__ ({ \
+  int16x8_t __s0_235 = __p0_235; \
+  int16x8_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int32x4_t __ret_235; \
+  __ret_235 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_235), __p1_235)); \
   __ret_235 = __builtin_shufflevector(__ret_235, __ret_235, 3, 2, 1, 0); \
   __ret_235; \
 })
 #endif
 
+#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
+  uint64_t __s0 = __p0; \
+  uint64_t __ret; \
+  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
+  __ret; \
+})
+#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
+  int64_t __s0 = __p0; \
+  int64_t __ret; \
+  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
+  __ret; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s8(__p0_236, __p1_236) __extension__ ({ \
-  int8x16_t __s0_236 = __p0_236; \
-  int16x8_t __ret_236; \
-  __ret_236 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_236), __p1_236)); \
+#define vshrn_high_n_u32(__p0_236, __p1_236, __p2_236) __extension__ ({ \
+  uint16x4_t __s0_236 = __p0_236; \
+  uint32x4_t __s1_236 = __p1_236; \
+  uint16x8_t __ret_236; \
+  __ret_236 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_236), (uint16x4_t)(vshrn_n_u32(__s1_236, __p2_236)))); \
   __ret_236; \
 })
 #else
-#define vshll_high_n_s8(__p0_237, __p1_237) __extension__ ({ \
-  int8x16_t __s0_237 = __p0_237; \
-  int8x16_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_237; \
-  __ret_237 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_237), __p1_237)); \
+#define vshrn_high_n_u32(__p0_237, __p1_237, __p2_237) __extension__ ({ \
+  uint16x4_t __s0_237 = __p0_237; \
+  uint32x4_t __s1_237 = __p1_237; \
+  uint16x4_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 3, 2, 1, 0); \
+  uint32x4_t __rev1_237;  __rev1_237 = __builtin_shufflevector(__s1_237, __s1_237, 3, 2, 1, 0); \
+  uint16x8_t __ret_237; \
+  __ret_237 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_237), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_237, __p2_237)))); \
   __ret_237 = __builtin_shufflevector(__ret_237, __ret_237, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_237; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s32(__p0_238, __p1_238) __extension__ ({ \
-  int32x4_t __s0_238 = __p0_238; \
-  int64x2_t __ret_238; \
-  __ret_238 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_238), __p1_238)); \
+#define vshrn_high_n_u64(__p0_238, __p1_238, __p2_238) __extension__ ({ \
+  uint32x2_t __s0_238 = __p0_238; \
+  uint64x2_t __s1_238 = __p1_238; \
+  uint32x4_t __ret_238; \
+  __ret_238 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_238), (uint32x2_t)(vshrn_n_u64(__s1_238, __p2_238)))); \
   __ret_238; \
 })
 #else
-#define vshll_high_n_s32(__p0_239, __p1_239) __extension__ ({ \
-  int32x4_t __s0_239 = __p0_239; \
-  int32x4_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 3, 2, 1, 0); \
-  int64x2_t __ret_239; \
-  __ret_239 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_239), __p1_239)); \
-  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 1, 0); \
+#define vshrn_high_n_u64(__p0_239, __p1_239, __p2_239) __extension__ ({ \
+  uint32x2_t __s0_239 = __p0_239; \
+  uint64x2_t __s1_239 = __p1_239; \
+  uint32x2_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 1, 0); \
+  uint64x2_t __rev1_239;  __rev1_239 = __builtin_shufflevector(__s1_239, __s1_239, 1, 0); \
+  uint32x4_t __ret_239; \
+  __ret_239 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_239), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_239, __p2_239)))); \
+  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 3, 2, 1, 0); \
   __ret_239; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s16(__p0_240, __p1_240) __extension__ ({ \
-  int16x8_t __s0_240 = __p0_240; \
-  int32x4_t __ret_240; \
-  __ret_240 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_240), __p1_240)); \
+#define vshrn_high_n_u16(__p0_240, __p1_240, __p2_240) __extension__ ({ \
+  uint8x8_t __s0_240 = __p0_240; \
+  uint16x8_t __s1_240 = __p1_240; \
+  uint8x16_t __ret_240; \
+  __ret_240 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_240), (uint8x8_t)(vshrn_n_u16(__s1_240, __p2_240)))); \
   __ret_240; \
 })
 #else
-#define vshll_high_n_s16(__p0_241, __p1_241) __extension__ ({ \
-  int16x8_t __s0_241 = __p0_241; \
-  int16x8_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret_241; \
-  __ret_241 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_241), __p1_241)); \
-  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 3, 2, 1, 0); \
+#define vshrn_high_n_u16(__p0_241, __p1_241, __p2_241) __extension__ ({ \
+  uint8x8_t __s0_241 = __p0_241; \
+  uint16x8_t __s1_241 = __p1_241; \
+  uint8x8_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint16x8_t __rev1_241;  __rev1_241 = __builtin_shufflevector(__s1_241, __s1_241, 7, 6, 5, 4, 3, 2, 1, 0); \
+  uint8x16_t __ret_241; \
+  __ret_241 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_241), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_241, __p2_241)))); \
+  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_241; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
-  uint16x4_t __s0_242 = __p0_242; \
-  uint32x4_t __s1_242 = __p1_242; \
-  uint16x8_t __ret_242; \
-  __ret_242 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_242), (uint16x4_t)(vshrn_n_u32(__s1_242, __p2_242)))); \
+#define vshrn_high_n_s32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
+  int16x4_t __s0_242 = __p0_242; \
+  int32x4_t __s1_242 = __p1_242; \
+  int16x8_t __ret_242; \
+  __ret_242 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_242), (int16x4_t)(vshrn_n_s32(__s1_242, __p2_242)))); \
   __ret_242; \
 })
 #else
-#define vshrn_high_n_u32(__p0_243, __p1_243, __p2_243) __extension__ ({ \
-  uint16x4_t __s0_243 = __p0_243; \
-  uint32x4_t __s1_243 = __p1_243; \
-  uint16x4_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 3, 2, 1, 0); \
-  uint32x4_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 3, 2, 1, 0); \
-  uint16x8_t __ret_243; \
-  __ret_243 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_243), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_243, __p2_243)))); \
+#define vshrn_high_n_s32(__p0_243, __p1_243, __p2_243) __extension__ ({ \
+  int16x4_t __s0_243 = __p0_243; \
+  int32x4_t __s1_243 = __p1_243; \
+  int16x4_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 3, 2, 1, 0); \
+  int32x4_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 3, 2, 1, 0); \
+  int16x8_t __ret_243; \
+  __ret_243 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_243), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_243, __p2_243)))); \
   __ret_243 = __builtin_shufflevector(__ret_243, __ret_243, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_243; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u64(__p0_244, __p1_244, __p2_244) __extension__ ({ \
-  uint32x2_t __s0_244 = __p0_244; \
-  uint64x2_t __s1_244 = __p1_244; \
-  uint32x4_t __ret_244; \
-  __ret_244 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_244), (uint32x2_t)(vshrn_n_u64(__s1_244, __p2_244)))); \
+#define vshrn_high_n_s64(__p0_244, __p1_244, __p2_244) __extension__ ({ \
+  int32x2_t __s0_244 = __p0_244; \
+  int64x2_t __s1_244 = __p1_244; \
+  int32x4_t __ret_244; \
+  __ret_244 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_244), (int32x2_t)(vshrn_n_s64(__s1_244, __p2_244)))); \
   __ret_244; \
 })
 #else
-#define vshrn_high_n_u64(__p0_245, __p1_245, __p2_245) __extension__ ({ \
-  uint32x2_t __s0_245 = __p0_245; \
-  uint64x2_t __s1_245 = __p1_245; \
-  uint32x2_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 1, 0); \
-  uint64x2_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 1, 0); \
-  uint32x4_t __ret_245; \
-  __ret_245 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_245), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_245, __p2_245)))); \
+#define vshrn_high_n_s64(__p0_245, __p1_245, __p2_245) __extension__ ({ \
+  int32x2_t __s0_245 = __p0_245; \
+  int64x2_t __s1_245 = __p1_245; \
+  int32x2_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 1, 0); \
+  int64x2_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 1, 0); \
+  int32x4_t __ret_245; \
+  __ret_245 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_245), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_245, __p2_245)))); \
   __ret_245 = __builtin_shufflevector(__ret_245, __ret_245, 3, 2, 1, 0); \
   __ret_245; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u16(__p0_246, __p1_246, __p2_246) __extension__ ({ \
-  uint8x8_t __s0_246 = __p0_246; \
-  uint16x8_t __s1_246 = __p1_246; \
-  uint8x16_t __ret_246; \
-  __ret_246 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_246), (uint8x8_t)(vshrn_n_u16(__s1_246, __p2_246)))); \
+#define vshrn_high_n_s16(__p0_246, __p1_246, __p2_246) __extension__ ({ \
+  int8x8_t __s0_246 = __p0_246; \
+  int16x8_t __s1_246 = __p1_246; \
+  int8x16_t __ret_246; \
+  __ret_246 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_246), (int8x8_t)(vshrn_n_s16(__s1_246, __p2_246)))); \
   __ret_246; \
 })
 #else
-#define vshrn_high_n_u16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
-  uint8x8_t __s0_247 = __p0_247; \
-  uint16x8_t __s1_247 = __p1_247; \
-  uint8x8_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_247; \
-  __ret_247 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_247), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_247, __p2_247)))); \
+#define vshrn_high_n_s16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
+  int8x8_t __s0_247 = __p0_247; \
+  int16x8_t __s1_247 = __p1_247; \
+  int8x8_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16x8_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int8x16_t __ret_247; \
+  __ret_247 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_247), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_247, __p2_247)))); \
   __ret_247 = __builtin_shufflevector(__ret_247, __ret_247, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
   __ret_247; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s32(__p0_248, __p1_248, __p2_248) __extension__ ({ \
-  int16x4_t __s0_248 = __p0_248; \
-  int32x4_t __s1_248 = __p1_248; \
-  int16x8_t __ret_248; \
-  __ret_248 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_248), (int16x4_t)(vshrn_n_s32(__s1_248, __p2_248)))); \
-  __ret_248; \
-})
-#else
-#define vshrn_high_n_s32(__p0_249, __p1_249, __p2_249) __extension__ ({ \
-  int16x4_t __s0_249 = __p0_249; \
-  int32x4_t __s1_249 = __p1_249; \
-  int16x4_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 3, 2, 1, 0); \
-  int32x4_t __rev1_249;  __rev1_249 = __builtin_shufflevector(__s1_249, __s1_249, 3, 2, 1, 0); \
-  int16x8_t __ret_249; \
-  __ret_249 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_249), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_249, __p2_249)))); \
-  __ret_249 = __builtin_shufflevector(__ret_249, __ret_249, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_249; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s64(__p0_250, __p1_250, __p2_250) __extension__ ({ \
-  int32x2_t __s0_250 = __p0_250; \
-  int64x2_t __s1_250 = __p1_250; \
-  int32x4_t __ret_250; \
-  __ret_250 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_250), (int32x2_t)(vshrn_n_s64(__s1_250, __p2_250)))); \
-  __ret_250; \
-})
-#else
-#define vshrn_high_n_s64(__p0_251, __p1_251, __p2_251) __extension__ ({ \
-  int32x2_t __s0_251 = __p0_251; \
-  int64x2_t __s1_251 = __p1_251; \
-  int32x2_t __rev0_251;  __rev0_251 = __builtin_shufflevector(__s0_251, __s0_251, 1, 0); \
-  int64x2_t __rev1_251;  __rev1_251 = __builtin_shufflevector(__s1_251, __s1_251, 1, 0); \
-  int32x4_t __ret_251; \
-  __ret_251 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_251), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_251, __p2_251)))); \
-  __ret_251 = __builtin_shufflevector(__ret_251, __ret_251, 3, 2, 1, 0); \
-  __ret_251; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s16(__p0_252, __p1_252, __p2_252) __extension__ ({ \
-  int8x8_t __s0_252 = __p0_252; \
-  int16x8_t __s1_252 = __p1_252; \
-  int8x16_t __ret_252; \
-  __ret_252 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_252), (int8x8_t)(vshrn_n_s16(__s1_252, __p2_252)))); \
-  __ret_252; \
-})
-#else
-#define vshrn_high_n_s16(__p0_253, __p1_253, __p2_253) __extension__ ({ \
-  int8x8_t __s0_253 = __p0_253; \
-  int16x8_t __s1_253 = __p1_253; \
-  int8x8_t __rev0_253;  __rev0_253 = __builtin_shufflevector(__s0_253, __s0_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_253;  __rev1_253 = __builtin_shufflevector(__s1_253, __s1_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_253; \
-  __ret_253 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_253), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_253, __p2_253)))); \
-  __ret_253 = __builtin_shufflevector(__ret_253, __ret_253, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_253; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66225,17 +55970,6 @@
   __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66243,17 +55977,6 @@
   __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -66261,16 +55984,6 @@
   __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -66292,62 +56005,26 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
   uint8_t __ret;
   __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
   uint32_t __ret;
   __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
   uint16_t __ret;
   __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
   uint8x16_t __ret;
@@ -66450,20 +56127,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
   uint16x4_t __ret;
@@ -66513,20 +56181,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vsqrt_f64(float64x1_t __p0) {
   float64x1_t __ret;
   __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
   return __ret;
 }
-#else
-__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float32x2_t vsqrt_f32(float32x2_t __p0) {
   float32x2_t __ret;
@@ -66543,7 +56202,6 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66551,17 +56209,6 @@
   __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66569,17 +56216,6 @@
   __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64_t __s0 = __p0; \
   uint64_t __s1 = __p1; \
@@ -66587,17 +56223,6 @@
   __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
   int64_t __s0 = __p0; \
   int64_t __s1 = __p1; \
@@ -66605,17 +56230,6 @@
   __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
   __ret; \
 })
-#else
-#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s0 = __p0; \
   poly64x1_t __s1 = __p1; \
@@ -66623,16 +56237,6 @@
   __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
   __ret; \
 })
-#else
-#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s0 = __p0; \
@@ -66654,18 +56258,10 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64(__p0, __p1) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
 })
-#else
-#define vst1_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64(__p0, __p1) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -66692,30 +56288,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64(__p0, __p1) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
 })
-#else
-#define vst1_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
 })
-#else
-#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x2_t __s1 = __p1; \
@@ -66742,30 +56322,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1_t __s1 = __p1; \
   __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
 })
-#else
-#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x2(__p0, __p1) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
 })
-#else
-#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
   poly64x2x2_t __s1 = __p1; \
@@ -66796,30 +56360,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x2(__p0, __p1) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
 })
-#else
-#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x3(__p0, __p1) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
 })
-#else
-#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
   poly64x2x3_t __s1 = __p1; \
@@ -66852,30 +56400,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x3(__p0, __p1) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
 })
-#else
-#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst1_p64_x4(__p0, __p1) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
 })
-#else
-#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
   poly64x2x4_t __s1 = __p1; \
@@ -66910,30 +56442,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst1_f64_x4(__p0, __p1) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
 })
-#else
-#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_p64(__p0, __p1) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
 })
-#else
-#define vst2_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x2_t __s1 = __p1; \
@@ -66994,30 +56510,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_f64(__p0, __p1) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
 })
-#else
-#define vst2_f64(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
 })
-#else
-#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x2_t __s1 = __p1; \
@@ -67123,54 +56623,22 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
 })
-#else
-#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
 })
-#else
-#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x2_t __s1 = __p1; \
   __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
 })
-#else
-#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_p64(__p0, __p1) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
 })
-#else
-#define vst3_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x3_t __s1 = __p1; \
@@ -67235,30 +56703,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_f64(__p0, __p1) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
 })
-#else
-#define vst3_f64(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
 })
-#else
-#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x3_t __s1 = __p1; \
@@ -67371,54 +56823,22 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
 })
-#else
-#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
 })
-#else
-#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x3_t __s1 = __p1; \
   __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
 })
-#else
-#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_p64(__p0, __p1) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
 })
-#else
-#define vst4_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4q_p64(__p0, __p1) __extension__ ({ \
   poly64x2x4_t __s1 = __p1; \
@@ -67487,30 +56907,14 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_f64(__p0, __p1) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
 })
-#else
-#define vst4_f64(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
   poly64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
 })
-#else
-#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-})
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 #define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
   poly8x16x4_t __s1 = __p1; \
@@ -67630,82 +57034,32 @@
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
   uint64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
 })
-#else
-#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
   float64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
 })
-#else
-#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
   int64x1x4_t __s1 = __p1; \
   __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
 })
-#else
-#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 #define vstrq_p128(__p0, __p1) __extension__ ({ \
   poly128_t __s1 = __p1; \
   __builtin_neon_vstrq_p128(__p0, __s1); \
 })
-#else
-#define vstrq_p128(__p0, __p1) __extension__ ({ \
-  poly128_t __s1 = __p1; \
-  __builtin_neon_vstrq_p128(__p0, __s1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
   float64x2_t __ret;
@@ -67723,20 +57077,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
   float64x1_t __ret;
   __ret = __p0 - __p1;
   return __ret;
 }
-#else
-__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
   uint16x8_t __ret;
@@ -68797,20 +58142,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
   uint64x2_t __ret;
@@ -68862,118 +58198,46 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
   uint64x1_t __ret;
   __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
   return __ret;
 }
-#else
-__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
   uint64_t __ret;
   __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
   return __ret;
 }
-#else
-__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
   int8_t __ret;
   __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
   return __ret;
 }
-#else
-__ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
   int32_t __ret;
   __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
   return __ret;
 }
-#else
-__ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
   int64_t __ret;
   __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
   return __ret;
 }
-#else
-__ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
   int16_t __ret;
   __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
   return __ret;
 }
-#else
-__ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
   int8x16_t __ret;
@@ -69076,20 +58340,11 @@
 }
 #endif
 
-#ifdef __LITTLE_ENDIAN__
 __ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
   int64x1_t __ret;
   __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
   return __ret;
 }
-#else
-__ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
 #ifdef __LITTLE_ENDIAN__
 __ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
   int16x4_t __ret;
@@ -71157,60 +60412,60 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vget_lane_f16(__p0_254, __p1_254) __extension__ ({ \
-  float16x4_t __s0_254 = __p0_254; \
-  float16_t __ret_254; \
-float16x4_t __reint_254 = __s0_254; \
-int16_t __reint1_254 = vget_lane_s16(*(int16x4_t *) &__reint_254, __p1_254); \
-  __ret_254 = *(float16_t *) &__reint1_254; \
-  __ret_254; \
+#define vget_lane_f16(__p0_248, __p1_248) __extension__ ({ \
+  float16x4_t __s0_248 = __p0_248; \
+  float16_t __ret_248; \
+float16x4_t __reint_248 = __s0_248; \
+int16_t __reint1_248 = vget_lane_s16(*(int16x4_t *) &__reint_248, __p1_248); \
+  __ret_248 = *(float16_t *) &__reint1_248; \
+  __ret_248; \
 })
 #else
-#define vget_lane_f16(__p0_255, __p1_255) __extension__ ({ \
-  float16x4_t __s0_255 = __p0_255; \
-  float16x4_t __rev0_255;  __rev0_255 = __builtin_shufflevector(__s0_255, __s0_255, 3, 2, 1, 0); \
-  float16_t __ret_255; \
-float16x4_t __reint_255 = __rev0_255; \
-int16_t __reint1_255 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_255, __p1_255); \
-  __ret_255 = *(float16_t *) &__reint1_255; \
-  __ret_255; \
+#define vget_lane_f16(__p0_249, __p1_249) __extension__ ({ \
+  float16x4_t __s0_249 = __p0_249; \
+  float16x4_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 3, 2, 1, 0); \
+  float16_t __ret_249; \
+float16x4_t __reint_249 = __rev0_249; \
+int16_t __reint1_249 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_249, __p1_249); \
+  __ret_249 = *(float16_t *) &__reint1_249; \
+  __ret_249; \
 })
-#define __noswap_vget_lane_f16(__p0_256, __p1_256) __extension__ ({ \
-  float16x4_t __s0_256 = __p0_256; \
-  float16_t __ret_256; \
-float16x4_t __reint_256 = __s0_256; \
-int16_t __reint1_256 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_256, __p1_256); \
-  __ret_256 = *(float16_t *) &__reint1_256; \
-  __ret_256; \
+#define __noswap_vget_lane_f16(__p0_250, __p1_250) __extension__ ({ \
+  float16x4_t __s0_250 = __p0_250; \
+  float16_t __ret_250; \
+float16x4_t __reint_250 = __s0_250; \
+int16_t __reint1_250 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_250, __p1_250); \
+  __ret_250 = *(float16_t *) &__reint1_250; \
+  __ret_250; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_f16(__p0_257, __p1_257) __extension__ ({ \
-  float16x8_t __s0_257 = __p0_257; \
-  float16_t __ret_257; \
-float16x8_t __reint_257 = __s0_257; \
-int16_t __reint1_257 = vgetq_lane_s16(*(int16x8_t *) &__reint_257, __p1_257); \
-  __ret_257 = *(float16_t *) &__reint1_257; \
-  __ret_257; \
+#define vgetq_lane_f16(__p0_251, __p1_251) __extension__ ({ \
+  float16x8_t __s0_251 = __p0_251; \
+  float16_t __ret_251; \
+float16x8_t __reint_251 = __s0_251; \
+int16_t __reint1_251 = vgetq_lane_s16(*(int16x8_t *) &__reint_251, __p1_251); \
+  __ret_251 = *(float16_t *) &__reint1_251; \
+  __ret_251; \
 })
 #else
-#define vgetq_lane_f16(__p0_258, __p1_258) __extension__ ({ \
-  float16x8_t __s0_258 = __p0_258; \
-  float16x8_t __rev0_258;  __rev0_258 = __builtin_shufflevector(__s0_258, __s0_258, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_258; \
-float16x8_t __reint_258 = __rev0_258; \
-int16_t __reint1_258 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_258, __p1_258); \
-  __ret_258 = *(float16_t *) &__reint1_258; \
-  __ret_258; \
+#define vgetq_lane_f16(__p0_252, __p1_252) __extension__ ({ \
+  float16x8_t __s0_252 = __p0_252; \
+  float16x8_t __rev0_252;  __rev0_252 = __builtin_shufflevector(__s0_252, __s0_252, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16_t __ret_252; \
+float16x8_t __reint_252 = __rev0_252; \
+int16_t __reint1_252 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_252, __p1_252); \
+  __ret_252 = *(float16_t *) &__reint1_252; \
+  __ret_252; \
 })
-#define __noswap_vgetq_lane_f16(__p0_259, __p1_259) __extension__ ({ \
-  float16x8_t __s0_259 = __p0_259; \
-  float16_t __ret_259; \
-float16x8_t __reint_259 = __s0_259; \
-int16_t __reint1_259 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_259, __p1_259); \
-  __ret_259 = *(float16_t *) &__reint1_259; \
-  __ret_259; \
+#define __noswap_vgetq_lane_f16(__p0_253, __p1_253) __extension__ ({ \
+  float16x8_t __s0_253 = __p0_253; \
+  float16_t __ret_253; \
+float16x8_t __reint_253 = __s0_253; \
+int16_t __reint1_253 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_253, __p1_253); \
+  __ret_253 = *(float16_t *) &__reint1_253; \
+  __ret_253; \
 })
 #endif
 
@@ -71859,708 +61114,672 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vset_lane_f16(__p0_260, __p1_260, __p2_260) __extension__ ({ \
-  float16_t __s0_260 = __p0_260; \
-  float16x4_t __s1_260 = __p1_260; \
-  float16x4_t __ret_260; \
-float16_t __reint_260 = __s0_260; \
-float16x4_t __reint1_260 = __s1_260; \
-int16x4_t __reint2_260 = vset_lane_s16(*(int16_t *) &__reint_260, *(int16x4_t *) &__reint1_260, __p2_260); \
-  __ret_260 = *(float16x4_t *) &__reint2_260; \
-  __ret_260; \
+#define vset_lane_f16(__p0_254, __p1_254, __p2_254) __extension__ ({ \
+  float16_t __s0_254 = __p0_254; \
+  float16x4_t __s1_254 = __p1_254; \
+  float16x4_t __ret_254; \
+float16_t __reint_254 = __s0_254; \
+float16x4_t __reint1_254 = __s1_254; \
+int16x4_t __reint2_254 = vset_lane_s16(*(int16_t *) &__reint_254, *(int16x4_t *) &__reint1_254, __p2_254); \
+  __ret_254 = *(float16x4_t *) &__reint2_254; \
+  __ret_254; \
 })
 #else
-#define vset_lane_f16(__p0_261, __p1_261, __p2_261) __extension__ ({ \
-  float16_t __s0_261 = __p0_261; \
-  float16x4_t __s1_261 = __p1_261; \
-  float16x4_t __rev1_261;  __rev1_261 = __builtin_shufflevector(__s1_261, __s1_261, 3, 2, 1, 0); \
-  float16x4_t __ret_261; \
-float16_t __reint_261 = __s0_261; \
-float16x4_t __reint1_261 = __rev1_261; \
-int16x4_t __reint2_261 = __noswap_vset_lane_s16(*(int16_t *) &__reint_261, *(int16x4_t *) &__reint1_261, __p2_261); \
-  __ret_261 = *(float16x4_t *) &__reint2_261; \
-  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 3, 2, 1, 0); \
-  __ret_261; \
+#define vset_lane_f16(__p0_255, __p1_255, __p2_255) __extension__ ({ \
+  float16_t __s0_255 = __p0_255; \
+  float16x4_t __s1_255 = __p1_255; \
+  float16x4_t __rev1_255;  __rev1_255 = __builtin_shufflevector(__s1_255, __s1_255, 3, 2, 1, 0); \
+  float16x4_t __ret_255; \
+float16_t __reint_255 = __s0_255; \
+float16x4_t __reint1_255 = __rev1_255; \
+int16x4_t __reint2_255 = __noswap_vset_lane_s16(*(int16_t *) &__reint_255, *(int16x4_t *) &__reint1_255, __p2_255); \
+  __ret_255 = *(float16x4_t *) &__reint2_255; \
+  __ret_255 = __builtin_shufflevector(__ret_255, __ret_255, 3, 2, 1, 0); \
+  __ret_255; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_f16(__p0_262, __p1_262, __p2_262) __extension__ ({ \
-  float16_t __s0_262 = __p0_262; \
-  float16x8_t __s1_262 = __p1_262; \
-  float16x8_t __ret_262; \
-float16_t __reint_262 = __s0_262; \
-float16x8_t __reint1_262 = __s1_262; \
-int16x8_t __reint2_262 = vsetq_lane_s16(*(int16_t *) &__reint_262, *(int16x8_t *) &__reint1_262, __p2_262); \
-  __ret_262 = *(float16x8_t *) &__reint2_262; \
-  __ret_262; \
+#define vsetq_lane_f16(__p0_256, __p1_256, __p2_256) __extension__ ({ \
+  float16_t __s0_256 = __p0_256; \
+  float16x8_t __s1_256 = __p1_256; \
+  float16x8_t __ret_256; \
+float16_t __reint_256 = __s0_256; \
+float16x8_t __reint1_256 = __s1_256; \
+int16x8_t __reint2_256 = vsetq_lane_s16(*(int16_t *) &__reint_256, *(int16x8_t *) &__reint1_256, __p2_256); \
+  __ret_256 = *(float16x8_t *) &__reint2_256; \
+  __ret_256; \
 })
 #else
-#define vsetq_lane_f16(__p0_263, __p1_263, __p2_263) __extension__ ({ \
-  float16_t __s0_263 = __p0_263; \
-  float16x8_t __s1_263 = __p1_263; \
-  float16x8_t __rev1_263;  __rev1_263 = __builtin_shufflevector(__s1_263, __s1_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret_263; \
-float16_t __reint_263 = __s0_263; \
-float16x8_t __reint1_263 = __rev1_263; \
-int16x8_t __reint2_263 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_263, *(int16x8_t *) &__reint1_263, __p2_263); \
-  __ret_263 = *(float16x8_t *) &__reint2_263; \
-  __ret_263 = __builtin_shufflevector(__ret_263, __ret_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_263; \
+#define vsetq_lane_f16(__p0_257, __p1_257, __p2_257) __extension__ ({ \
+  float16_t __s0_257 = __p0_257; \
+  float16x8_t __s1_257 = __p1_257; \
+  float16x8_t __rev1_257;  __rev1_257 = __builtin_shufflevector(__s1_257, __s1_257, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __ret_257; \
+float16_t __reint_257 = __s0_257; \
+float16x8_t __reint1_257 = __rev1_257; \
+int16x8_t __reint2_257 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_257, *(int16x8_t *) &__reint1_257, __p2_257); \
+  __ret_257 = *(float16x8_t *) &__reint2_257; \
+  __ret_257 = __builtin_shufflevector(__ret_257, __ret_257, 7, 6, 5, 4, 3, 2, 1, 0); \
+  __ret_257; \
 })
 #endif
 
 #if defined(__ARM_FEATURE_FP16FML) && defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_high_f16(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
-  float32x4_t __s0_264 = __p0_264; \
-  float16x8_t __s1_264 = __p1_264; \
+#define vfmlalq_lane_high_f16(__p0_258, __p1_258, __p2_258, __p3_258) __extension__ ({ \
+  float32x4_t __s0_258 = __p0_258; \
+  float16x8_t __s1_258 = __p1_258; \
+  float16x4_t __s2_258 = __p2_258; \
+  float32x4_t __ret_258; \
+  __ret_258 = vfmlalq_high_f16(__s0_258, __s1_258, (float16x8_t) {vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258), vget_lane_f16(__s2_258, __p3_258)}); \
+  __ret_258; \
+})
+#else
+#define vfmlalq_lane_high_f16(__p0_259, __p1_259, __p2_259, __p3_259) __extension__ ({ \
+  float32x4_t __s0_259 = __p0_259; \
+  float16x8_t __s1_259 = __p1_259; \
+  float16x4_t __s2_259 = __p2_259; \
+  float32x4_t __rev0_259;  __rev0_259 = __builtin_shufflevector(__s0_259, __s0_259, 3, 2, 1, 0); \
+  float16x8_t __rev1_259;  __rev1_259 = __builtin_shufflevector(__s1_259, __s1_259, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_259;  __rev2_259 = __builtin_shufflevector(__s2_259, __s2_259, 3, 2, 1, 0); \
+  float32x4_t __ret_259; \
+  __ret_259 = __noswap_vfmlalq_high_f16(__rev0_259, __rev1_259, (float16x8_t) {__noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259), __noswap_vget_lane_f16(__rev2_259, __p3_259)}); \
+  __ret_259 = __builtin_shufflevector(__ret_259, __ret_259, 3, 2, 1, 0); \
+  __ret_259; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlal_lane_high_f16(__p0_260, __p1_260, __p2_260, __p3_260) __extension__ ({ \
+  float32x2_t __s0_260 = __p0_260; \
+  float16x4_t __s1_260 = __p1_260; \
+  float16x4_t __s2_260 = __p2_260; \
+  float32x2_t __ret_260; \
+  __ret_260 = vfmlal_high_f16(__s0_260, __s1_260, (float16x4_t) {vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260), vget_lane_f16(__s2_260, __p3_260)}); \
+  __ret_260; \
+})
+#else
+#define vfmlal_lane_high_f16(__p0_261, __p1_261, __p2_261, __p3_261) __extension__ ({ \
+  float32x2_t __s0_261 = __p0_261; \
+  float16x4_t __s1_261 = __p1_261; \
+  float16x4_t __s2_261 = __p2_261; \
+  float32x2_t __rev0_261;  __rev0_261 = __builtin_shufflevector(__s0_261, __s0_261, 1, 0); \
+  float16x4_t __rev1_261;  __rev1_261 = __builtin_shufflevector(__s1_261, __s1_261, 3, 2, 1, 0); \
+  float16x4_t __rev2_261;  __rev2_261 = __builtin_shufflevector(__s2_261, __s2_261, 3, 2, 1, 0); \
+  float32x2_t __ret_261; \
+  __ret_261 = __noswap_vfmlal_high_f16(__rev0_261, __rev1_261, (float16x4_t) {__noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261), __noswap_vget_lane_f16(__rev2_261, __p3_261)}); \
+  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 1, 0); \
+  __ret_261; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlalq_lane_low_f16(__p0_262, __p1_262, __p2_262, __p3_262) __extension__ ({ \
+  float32x4_t __s0_262 = __p0_262; \
+  float16x8_t __s1_262 = __p1_262; \
+  float16x4_t __s2_262 = __p2_262; \
+  float32x4_t __ret_262; \
+  __ret_262 = vfmlalq_low_f16(__s0_262, __s1_262, (float16x8_t) {vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262), vget_lane_f16(__s2_262, __p3_262)}); \
+  __ret_262; \
+})
+#else
+#define vfmlalq_lane_low_f16(__p0_263, __p1_263, __p2_263, __p3_263) __extension__ ({ \
+  float32x4_t __s0_263 = __p0_263; \
+  float16x8_t __s1_263 = __p1_263; \
+  float16x4_t __s2_263 = __p2_263; \
+  float32x4_t __rev0_263;  __rev0_263 = __builtin_shufflevector(__s0_263, __s0_263, 3, 2, 1, 0); \
+  float16x8_t __rev1_263;  __rev1_263 = __builtin_shufflevector(__s1_263, __s1_263, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_263;  __rev2_263 = __builtin_shufflevector(__s2_263, __s2_263, 3, 2, 1, 0); \
+  float32x4_t __ret_263; \
+  __ret_263 = __noswap_vfmlalq_low_f16(__rev0_263, __rev1_263, (float16x8_t) {__noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263), __noswap_vget_lane_f16(__rev2_263, __p3_263)}); \
+  __ret_263 = __builtin_shufflevector(__ret_263, __ret_263, 3, 2, 1, 0); \
+  __ret_263; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vfmlal_lane_low_f16(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
+  float32x2_t __s0_264 = __p0_264; \
+  float16x4_t __s1_264 = __p1_264; \
   float16x4_t __s2_264 = __p2_264; \
-  float32x4_t __ret_264; \
-  __ret_264 = vfmlalq_high_f16(__s0_264, __s1_264, (float16x8_t) {vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264)}); \
+  float32x2_t __ret_264; \
+  __ret_264 = vfmlal_low_f16(__s0_264, __s1_264, (float16x4_t) {vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264)}); \
   __ret_264; \
 })
 #else
-#define vfmlalq_lane_high_f16(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
-  float32x4_t __s0_265 = __p0_265; \
-  float16x8_t __s1_265 = __p1_265; \
+#define vfmlal_lane_low_f16(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
+  float32x2_t __s0_265 = __p0_265; \
+  float16x4_t __s1_265 = __p1_265; \
   float16x4_t __s2_265 = __p2_265; \
-  float32x4_t __rev0_265;  __rev0_265 = __builtin_shufflevector(__s0_265, __s0_265, 3, 2, 1, 0); \
-  float16x8_t __rev1_265;  __rev1_265 = __builtin_shufflevector(__s1_265, __s1_265, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_265;  __rev0_265 = __builtin_shufflevector(__s0_265, __s0_265, 1, 0); \
+  float16x4_t __rev1_265;  __rev1_265 = __builtin_shufflevector(__s1_265, __s1_265, 3, 2, 1, 0); \
   float16x4_t __rev2_265;  __rev2_265 = __builtin_shufflevector(__s2_265, __s2_265, 3, 2, 1, 0); \
-  float32x4_t __ret_265; \
-  __ret_265 = __noswap_vfmlalq_high_f16(__rev0_265, __rev1_265, (float16x8_t) {__noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265)}); \
-  __ret_265 = __builtin_shufflevector(__ret_265, __ret_265, 3, 2, 1, 0); \
+  float32x2_t __ret_265; \
+  __ret_265 = __noswap_vfmlal_low_f16(__rev0_265, __rev1_265, (float16x4_t) {__noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265)}); \
+  __ret_265 = __builtin_shufflevector(__ret_265, __ret_265, 1, 0); \
   __ret_265; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_high_f16(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
-  float32x2_t __s0_266 = __p0_266; \
-  float16x4_t __s1_266 = __p1_266; \
-  float16x4_t __s2_266 = __p2_266; \
-  float32x2_t __ret_266; \
-  __ret_266 = vfmlal_high_f16(__s0_266, __s1_266, (float16x4_t) {vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266)}); \
+#define vfmlalq_laneq_high_f16(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
+  float32x4_t __s0_266 = __p0_266; \
+  float16x8_t __s1_266 = __p1_266; \
+  float16x8_t __s2_266 = __p2_266; \
+  float32x4_t __ret_266; \
+  __ret_266 = vfmlalq_high_f16(__s0_266, __s1_266, (float16x8_t) {vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266), vgetq_lane_f16(__s2_266, __p3_266)}); \
   __ret_266; \
 })
 #else
-#define vfmlal_lane_high_f16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
-  float32x2_t __s0_267 = __p0_267; \
-  float16x4_t __s1_267 = __p1_267; \
-  float16x4_t __s2_267 = __p2_267; \
-  float32x2_t __rev0_267;  __rev0_267 = __builtin_shufflevector(__s0_267, __s0_267, 1, 0); \
-  float16x4_t __rev1_267;  __rev1_267 = __builtin_shufflevector(__s1_267, __s1_267, 3, 2, 1, 0); \
-  float16x4_t __rev2_267;  __rev2_267 = __builtin_shufflevector(__s2_267, __s2_267, 3, 2, 1, 0); \
-  float32x2_t __ret_267; \
-  __ret_267 = __noswap_vfmlal_high_f16(__rev0_267, __rev1_267, (float16x4_t) {__noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267)}); \
-  __ret_267 = __builtin_shufflevector(__ret_267, __ret_267, 1, 0); \
+#define vfmlalq_laneq_high_f16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
+  float32x4_t __s0_267 = __p0_267; \
+  float16x8_t __s1_267 = __p1_267; \
+  float16x8_t __s2_267 = __p2_267; \
+  float32x4_t __rev0_267;  __rev0_267 = __builtin_shufflevector(__s0_267, __s0_267, 3, 2, 1, 0); \
+  float16x8_t __rev1_267;  __rev1_267 = __builtin_shufflevector(__s1_267, __s1_267, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_267;  __rev2_267 = __builtin_shufflevector(__s2_267, __s2_267, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_267; \
+  __ret_267 = __noswap_vfmlalq_high_f16(__rev0_267, __rev1_267, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267), __noswap_vgetq_lane_f16(__rev2_267, __p3_267)}); \
+  __ret_267 = __builtin_shufflevector(__ret_267, __ret_267, 3, 2, 1, 0); \
   __ret_267; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_low_f16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
-  float32x4_t __s0_268 = __p0_268; \
-  float16x8_t __s1_268 = __p1_268; \
-  float16x4_t __s2_268 = __p2_268; \
-  float32x4_t __ret_268; \
-  __ret_268 = vfmlalq_low_f16(__s0_268, __s1_268, (float16x8_t) {vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268)}); \
+#define vfmlal_laneq_high_f16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
+  float32x2_t __s0_268 = __p0_268; \
+  float16x4_t __s1_268 = __p1_268; \
+  float16x8_t __s2_268 = __p2_268; \
+  float32x2_t __ret_268; \
+  __ret_268 = vfmlal_high_f16(__s0_268, __s1_268, (float16x4_t) {vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268), vgetq_lane_f16(__s2_268, __p3_268)}); \
   __ret_268; \
 })
 #else
-#define vfmlalq_lane_low_f16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
-  float32x4_t __s0_269 = __p0_269; \
-  float16x8_t __s1_269 = __p1_269; \
-  float16x4_t __s2_269 = __p2_269; \
-  float32x4_t __rev0_269;  __rev0_269 = __builtin_shufflevector(__s0_269, __s0_269, 3, 2, 1, 0); \
-  float16x8_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_269;  __rev2_269 = __builtin_shufflevector(__s2_269, __s2_269, 3, 2, 1, 0); \
-  float32x4_t __ret_269; \
-  __ret_269 = __noswap_vfmlalq_low_f16(__rev0_269, __rev1_269, (float16x8_t) {__noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269)}); \
-  __ret_269 = __builtin_shufflevector(__ret_269, __ret_269, 3, 2, 1, 0); \
+#define vfmlal_laneq_high_f16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
+  float32x2_t __s0_269 = __p0_269; \
+  float16x4_t __s1_269 = __p1_269; \
+  float16x8_t __s2_269 = __p2_269; \
+  float32x2_t __rev0_269;  __rev0_269 = __builtin_shufflevector(__s0_269, __s0_269, 1, 0); \
+  float16x4_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 3, 2, 1, 0); \
+  float16x8_t __rev2_269;  __rev2_269 = __builtin_shufflevector(__s2_269, __s2_269, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __ret_269; \
+  __ret_269 = __noswap_vfmlal_high_f16(__rev0_269, __rev1_269, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269), __noswap_vgetq_lane_f16(__rev2_269, __p3_269)}); \
+  __ret_269 = __builtin_shufflevector(__ret_269, __ret_269, 1, 0); \
   __ret_269; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_low_f16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
-  float32x2_t __s0_270 = __p0_270; \
-  float16x4_t __s1_270 = __p1_270; \
-  float16x4_t __s2_270 = __p2_270; \
-  float32x2_t __ret_270; \
-  __ret_270 = vfmlal_low_f16(__s0_270, __s1_270, (float16x4_t) {vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270)}); \
+#define vfmlalq_laneq_low_f16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
+  float32x4_t __s0_270 = __p0_270; \
+  float16x8_t __s1_270 = __p1_270; \
+  float16x8_t __s2_270 = __p2_270; \
+  float32x4_t __ret_270; \
+  __ret_270 = vfmlalq_low_f16(__s0_270, __s1_270, (float16x8_t) {vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270), vgetq_lane_f16(__s2_270, __p3_270)}); \
   __ret_270; \
 })
 #else
-#define vfmlal_lane_low_f16(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
-  float32x2_t __s0_271 = __p0_271; \
-  float16x4_t __s1_271 = __p1_271; \
-  float16x4_t __s2_271 = __p2_271; \
-  float32x2_t __rev0_271;  __rev0_271 = __builtin_shufflevector(__s0_271, __s0_271, 1, 0); \
-  float16x4_t __rev1_271;  __rev1_271 = __builtin_shufflevector(__s1_271, __s1_271, 3, 2, 1, 0); \
-  float16x4_t __rev2_271;  __rev2_271 = __builtin_shufflevector(__s2_271, __s2_271, 3, 2, 1, 0); \
-  float32x2_t __ret_271; \
-  __ret_271 = __noswap_vfmlal_low_f16(__rev0_271, __rev1_271, (float16x4_t) {__noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271)}); \
-  __ret_271 = __builtin_shufflevector(__ret_271, __ret_271, 1, 0); \
+#define vfmlalq_laneq_low_f16(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
+  float32x4_t __s0_271 = __p0_271; \
+  float16x8_t __s1_271 = __p1_271; \
+  float16x8_t __s2_271 = __p2_271; \
+  float32x4_t __rev0_271;  __rev0_271 = __builtin_shufflevector(__s0_271, __s0_271, 3, 2, 1, 0); \
+  float16x8_t __rev1_271;  __rev1_271 = __builtin_shufflevector(__s1_271, __s1_271, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_271;  __rev2_271 = __builtin_shufflevector(__s2_271, __s2_271, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_271; \
+  __ret_271 = __noswap_vfmlalq_low_f16(__rev0_271, __rev1_271, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271), __noswap_vgetq_lane_f16(__rev2_271, __p3_271)}); \
+  __ret_271 = __builtin_shufflevector(__ret_271, __ret_271, 3, 2, 1, 0); \
   __ret_271; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_high_f16(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
-  float32x4_t __s0_272 = __p0_272; \
-  float16x8_t __s1_272 = __p1_272; \
+#define vfmlal_laneq_low_f16(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
+  float32x2_t __s0_272 = __p0_272; \
+  float16x4_t __s1_272 = __p1_272; \
   float16x8_t __s2_272 = __p2_272; \
-  float32x4_t __ret_272; \
-  __ret_272 = vfmlalq_high_f16(__s0_272, __s1_272, (float16x8_t) {vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272)}); \
+  float32x2_t __ret_272; \
+  __ret_272 = vfmlal_low_f16(__s0_272, __s1_272, (float16x4_t) {vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272)}); \
   __ret_272; \
 })
 #else
-#define vfmlalq_laneq_high_f16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
-  float32x4_t __s0_273 = __p0_273; \
-  float16x8_t __s1_273 = __p1_273; \
+#define vfmlal_laneq_low_f16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
+  float32x2_t __s0_273 = __p0_273; \
+  float16x4_t __s1_273 = __p1_273; \
   float16x8_t __s2_273 = __p2_273; \
-  float32x4_t __rev0_273;  __rev0_273 = __builtin_shufflevector(__s0_273, __s0_273, 3, 2, 1, 0); \
-  float16x8_t __rev1_273;  __rev1_273 = __builtin_shufflevector(__s1_273, __s1_273, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_273;  __rev0_273 = __builtin_shufflevector(__s0_273, __s0_273, 1, 0); \
+  float16x4_t __rev1_273;  __rev1_273 = __builtin_shufflevector(__s1_273, __s1_273, 3, 2, 1, 0); \
   float16x8_t __rev2_273;  __rev2_273 = __builtin_shufflevector(__s2_273, __s2_273, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_273; \
-  __ret_273 = __noswap_vfmlalq_high_f16(__rev0_273, __rev1_273, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273)}); \
-  __ret_273 = __builtin_shufflevector(__ret_273, __ret_273, 3, 2, 1, 0); \
+  float32x2_t __ret_273; \
+  __ret_273 = __noswap_vfmlal_low_f16(__rev0_273, __rev1_273, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273)}); \
+  __ret_273 = __builtin_shufflevector(__ret_273, __ret_273, 1, 0); \
   __ret_273; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_high_f16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
-  float32x2_t __s0_274 = __p0_274; \
-  float16x4_t __s1_274 = __p1_274; \
-  float16x8_t __s2_274 = __p2_274; \
-  float32x2_t __ret_274; \
-  __ret_274 = vfmlal_high_f16(__s0_274, __s1_274, (float16x4_t) {vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274)}); \
+#define vfmlslq_lane_high_f16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
+  float32x4_t __s0_274 = __p0_274; \
+  float16x8_t __s1_274 = __p1_274; \
+  float16x4_t __s2_274 = __p2_274; \
+  float32x4_t __ret_274; \
+  __ret_274 = vfmlslq_high_f16(__s0_274, __s1_274, (float16x8_t) {vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274), vget_lane_f16(__s2_274, __p3_274)}); \
   __ret_274; \
 })
 #else
-#define vfmlal_laneq_high_f16(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
-  float32x2_t __s0_275 = __p0_275; \
-  float16x4_t __s1_275 = __p1_275; \
-  float16x8_t __s2_275 = __p2_275; \
-  float32x2_t __rev0_275;  __rev0_275 = __builtin_shufflevector(__s0_275, __s0_275, 1, 0); \
-  float16x4_t __rev1_275;  __rev1_275 = __builtin_shufflevector(__s1_275, __s1_275, 3, 2, 1, 0); \
-  float16x8_t __rev2_275;  __rev2_275 = __builtin_shufflevector(__s2_275, __s2_275, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_275; \
-  __ret_275 = __noswap_vfmlal_high_f16(__rev0_275, __rev1_275, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275)}); \
-  __ret_275 = __builtin_shufflevector(__ret_275, __ret_275, 1, 0); \
+#define vfmlslq_lane_high_f16(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
+  float32x4_t __s0_275 = __p0_275; \
+  float16x8_t __s1_275 = __p1_275; \
+  float16x4_t __s2_275 = __p2_275; \
+  float32x4_t __rev0_275;  __rev0_275 = __builtin_shufflevector(__s0_275, __s0_275, 3, 2, 1, 0); \
+  float16x8_t __rev1_275;  __rev1_275 = __builtin_shufflevector(__s1_275, __s1_275, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_275;  __rev2_275 = __builtin_shufflevector(__s2_275, __s2_275, 3, 2, 1, 0); \
+  float32x4_t __ret_275; \
+  __ret_275 = __noswap_vfmlslq_high_f16(__rev0_275, __rev1_275, (float16x8_t) {__noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275), __noswap_vget_lane_f16(__rev2_275, __p3_275)}); \
+  __ret_275 = __builtin_shufflevector(__ret_275, __ret_275, 3, 2, 1, 0); \
   __ret_275; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_low_f16(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
-  float32x4_t __s0_276 = __p0_276; \
-  float16x8_t __s1_276 = __p1_276; \
-  float16x8_t __s2_276 = __p2_276; \
-  float32x4_t __ret_276; \
-  __ret_276 = vfmlalq_low_f16(__s0_276, __s1_276, (float16x8_t) {vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276)}); \
+#define vfmlsl_lane_high_f16(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
+  float32x2_t __s0_276 = __p0_276; \
+  float16x4_t __s1_276 = __p1_276; \
+  float16x4_t __s2_276 = __p2_276; \
+  float32x2_t __ret_276; \
+  __ret_276 = vfmlsl_high_f16(__s0_276, __s1_276, (float16x4_t) {vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276), vget_lane_f16(__s2_276, __p3_276)}); \
   __ret_276; \
 })
 #else
-#define vfmlalq_laneq_low_f16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
-  float32x4_t __s0_277 = __p0_277; \
-  float16x8_t __s1_277 = __p1_277; \
-  float16x8_t __s2_277 = __p2_277; \
-  float32x4_t __rev0_277;  __rev0_277 = __builtin_shufflevector(__s0_277, __s0_277, 3, 2, 1, 0); \
-  float16x8_t __rev1_277;  __rev1_277 = __builtin_shufflevector(__s1_277, __s1_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_277;  __rev2_277 = __builtin_shufflevector(__s2_277, __s2_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_277; \
-  __ret_277 = __noswap_vfmlalq_low_f16(__rev0_277, __rev1_277, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277)}); \
-  __ret_277 = __builtin_shufflevector(__ret_277, __ret_277, 3, 2, 1, 0); \
+#define vfmlsl_lane_high_f16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
+  float32x2_t __s0_277 = __p0_277; \
+  float16x4_t __s1_277 = __p1_277; \
+  float16x4_t __s2_277 = __p2_277; \
+  float32x2_t __rev0_277;  __rev0_277 = __builtin_shufflevector(__s0_277, __s0_277, 1, 0); \
+  float16x4_t __rev1_277;  __rev1_277 = __builtin_shufflevector(__s1_277, __s1_277, 3, 2, 1, 0); \
+  float16x4_t __rev2_277;  __rev2_277 = __builtin_shufflevector(__s2_277, __s2_277, 3, 2, 1, 0); \
+  float32x2_t __ret_277; \
+  __ret_277 = __noswap_vfmlsl_high_f16(__rev0_277, __rev1_277, (float16x4_t) {__noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277), __noswap_vget_lane_f16(__rev2_277, __p3_277)}); \
+  __ret_277 = __builtin_shufflevector(__ret_277, __ret_277, 1, 0); \
   __ret_277; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_low_f16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
-  float32x2_t __s0_278 = __p0_278; \
-  float16x4_t __s1_278 = __p1_278; \
-  float16x8_t __s2_278 = __p2_278; \
-  float32x2_t __ret_278; \
-  __ret_278 = vfmlal_low_f16(__s0_278, __s1_278, (float16x4_t) {vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278)}); \
+#define vfmlslq_lane_low_f16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
+  float32x4_t __s0_278 = __p0_278; \
+  float16x8_t __s1_278 = __p1_278; \
+  float16x4_t __s2_278 = __p2_278; \
+  float32x4_t __ret_278; \
+  __ret_278 = vfmlslq_low_f16(__s0_278, __s1_278, (float16x8_t) {vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278), vget_lane_f16(__s2_278, __p3_278)}); \
   __ret_278; \
 })
 #else
-#define vfmlal_laneq_low_f16(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
-  float32x2_t __s0_279 = __p0_279; \
-  float16x4_t __s1_279 = __p1_279; \
-  float16x8_t __s2_279 = __p2_279; \
-  float32x2_t __rev0_279;  __rev0_279 = __builtin_shufflevector(__s0_279, __s0_279, 1, 0); \
-  float16x4_t __rev1_279;  __rev1_279 = __builtin_shufflevector(__s1_279, __s1_279, 3, 2, 1, 0); \
-  float16x8_t __rev2_279;  __rev2_279 = __builtin_shufflevector(__s2_279, __s2_279, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_279; \
-  __ret_279 = __noswap_vfmlal_low_f16(__rev0_279, __rev1_279, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279)}); \
-  __ret_279 = __builtin_shufflevector(__ret_279, __ret_279, 1, 0); \
+#define vfmlslq_lane_low_f16(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
+  float32x4_t __s0_279 = __p0_279; \
+  float16x8_t __s1_279 = __p1_279; \
+  float16x4_t __s2_279 = __p2_279; \
+  float32x4_t __rev0_279;  __rev0_279 = __builtin_shufflevector(__s0_279, __s0_279, 3, 2, 1, 0); \
+  float16x8_t __rev1_279;  __rev1_279 = __builtin_shufflevector(__s1_279, __s1_279, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x4_t __rev2_279;  __rev2_279 = __builtin_shufflevector(__s2_279, __s2_279, 3, 2, 1, 0); \
+  float32x4_t __ret_279; \
+  __ret_279 = __noswap_vfmlslq_low_f16(__rev0_279, __rev1_279, (float16x8_t) {__noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279), __noswap_vget_lane_f16(__rev2_279, __p3_279)}); \
+  __ret_279 = __builtin_shufflevector(__ret_279, __ret_279, 3, 2, 1, 0); \
   __ret_279; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_high_f16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
-  float32x4_t __s0_280 = __p0_280; \
-  float16x8_t __s1_280 = __p1_280; \
+#define vfmlsl_lane_low_f16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
+  float32x2_t __s0_280 = __p0_280; \
+  float16x4_t __s1_280 = __p1_280; \
   float16x4_t __s2_280 = __p2_280; \
-  float32x4_t __ret_280; \
-  __ret_280 = vfmlslq_high_f16(__s0_280, __s1_280, (float16x8_t) {vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280)}); \
+  float32x2_t __ret_280; \
+  __ret_280 = vfmlsl_low_f16(__s0_280, __s1_280, (float16x4_t) {vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280)}); \
   __ret_280; \
 })
 #else
-#define vfmlslq_lane_high_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
-  float32x4_t __s0_281 = __p0_281; \
-  float16x8_t __s1_281 = __p1_281; \
+#define vfmlsl_lane_low_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
+  float32x2_t __s0_281 = __p0_281; \
+  float16x4_t __s1_281 = __p1_281; \
   float16x4_t __s2_281 = __p2_281; \
-  float32x4_t __rev0_281;  __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 3, 2, 1, 0); \
-  float16x8_t __rev1_281;  __rev1_281 = __builtin_shufflevector(__s1_281, __s1_281, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_281;  __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 1, 0); \
+  float16x4_t __rev1_281;  __rev1_281 = __builtin_shufflevector(__s1_281, __s1_281, 3, 2, 1, 0); \
   float16x4_t __rev2_281;  __rev2_281 = __builtin_shufflevector(__s2_281, __s2_281, 3, 2, 1, 0); \
-  float32x4_t __ret_281; \
-  __ret_281 = __noswap_vfmlslq_high_f16(__rev0_281, __rev1_281, (float16x8_t) {__noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281)}); \
-  __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 3, 2, 1, 0); \
+  float32x2_t __ret_281; \
+  __ret_281 = __noswap_vfmlsl_low_f16(__rev0_281, __rev1_281, (float16x4_t) {__noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281)}); \
+  __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 1, 0); \
   __ret_281; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_high_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
-  float32x2_t __s0_282 = __p0_282; \
-  float16x4_t __s1_282 = __p1_282; \
-  float16x4_t __s2_282 = __p2_282; \
-  float32x2_t __ret_282; \
-  __ret_282 = vfmlsl_high_f16(__s0_282, __s1_282, (float16x4_t) {vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282)}); \
+#define vfmlslq_laneq_high_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
+  float32x4_t __s0_282 = __p0_282; \
+  float16x8_t __s1_282 = __p1_282; \
+  float16x8_t __s2_282 = __p2_282; \
+  float32x4_t __ret_282; \
+  __ret_282 = vfmlslq_high_f16(__s0_282, __s1_282, (float16x8_t) {vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282), vgetq_lane_f16(__s2_282, __p3_282)}); \
   __ret_282; \
 })
 #else
-#define vfmlsl_lane_high_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
-  float32x2_t __s0_283 = __p0_283; \
-  float16x4_t __s1_283 = __p1_283; \
-  float16x4_t __s2_283 = __p2_283; \
-  float32x2_t __rev0_283;  __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 1, 0); \
-  float16x4_t __rev1_283;  __rev1_283 = __builtin_shufflevector(__s1_283, __s1_283, 3, 2, 1, 0); \
-  float16x4_t __rev2_283;  __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 3, 2, 1, 0); \
-  float32x2_t __ret_283; \
-  __ret_283 = __noswap_vfmlsl_high_f16(__rev0_283, __rev1_283, (float16x4_t) {__noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283)}); \
-  __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 1, 0); \
+#define vfmlslq_laneq_high_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
+  float32x4_t __s0_283 = __p0_283; \
+  float16x8_t __s1_283 = __p1_283; \
+  float16x8_t __s2_283 = __p2_283; \
+  float32x4_t __rev0_283;  __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 3, 2, 1, 0); \
+  float16x8_t __rev1_283;  __rev1_283 = __builtin_shufflevector(__s1_283, __s1_283, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_283;  __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_283; \
+  __ret_283 = __noswap_vfmlslq_high_f16(__rev0_283, __rev1_283, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283), __noswap_vgetq_lane_f16(__rev2_283, __p3_283)}); \
+  __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 3, 2, 1, 0); \
   __ret_283; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_low_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
-  float32x4_t __s0_284 = __p0_284; \
-  float16x8_t __s1_284 = __p1_284; \
-  float16x4_t __s2_284 = __p2_284; \
-  float32x4_t __ret_284; \
-  __ret_284 = vfmlslq_low_f16(__s0_284, __s1_284, (float16x8_t) {vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284)}); \
+#define vfmlsl_laneq_high_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
+  float32x2_t __s0_284 = __p0_284; \
+  float16x4_t __s1_284 = __p1_284; \
+  float16x8_t __s2_284 = __p2_284; \
+  float32x2_t __ret_284; \
+  __ret_284 = vfmlsl_high_f16(__s0_284, __s1_284, (float16x4_t) {vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284), vgetq_lane_f16(__s2_284, __p3_284)}); \
   __ret_284; \
 })
 #else
-#define vfmlslq_lane_low_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
-  float32x4_t __s0_285 = __p0_285; \
-  float16x8_t __s1_285 = __p1_285; \
-  float16x4_t __s2_285 = __p2_285; \
-  float32x4_t __rev0_285;  __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 3, 2, 1, 0); \
-  float16x8_t __rev1_285;  __rev1_285 = __builtin_shufflevector(__s1_285, __s1_285, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_285;  __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 3, 2, 1, 0); \
-  float32x4_t __ret_285; \
-  __ret_285 = __noswap_vfmlslq_low_f16(__rev0_285, __rev1_285, (float16x8_t) {__noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285)}); \
-  __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 3, 2, 1, 0); \
+#define vfmlsl_laneq_high_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
+  float32x2_t __s0_285 = __p0_285; \
+  float16x4_t __s1_285 = __p1_285; \
+  float16x8_t __s2_285 = __p2_285; \
+  float32x2_t __rev0_285;  __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 1, 0); \
+  float16x4_t __rev1_285;  __rev1_285 = __builtin_shufflevector(__s1_285, __s1_285, 3, 2, 1, 0); \
+  float16x8_t __rev2_285;  __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __ret_285; \
+  __ret_285 = __noswap_vfmlsl_high_f16(__rev0_285, __rev1_285, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285), __noswap_vgetq_lane_f16(__rev2_285, __p3_285)}); \
+  __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 1, 0); \
   __ret_285; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_low_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
-  float32x2_t __s0_286 = __p0_286; \
-  float16x4_t __s1_286 = __p1_286; \
-  float16x4_t __s2_286 = __p2_286; \
-  float32x2_t __ret_286; \
-  __ret_286 = vfmlsl_low_f16(__s0_286, __s1_286, (float16x4_t) {vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286)}); \
+#define vfmlslq_laneq_low_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
+  float32x4_t __s0_286 = __p0_286; \
+  float16x8_t __s1_286 = __p1_286; \
+  float16x8_t __s2_286 = __p2_286; \
+  float32x4_t __ret_286; \
+  __ret_286 = vfmlslq_low_f16(__s0_286, __s1_286, (float16x8_t) {vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286), vgetq_lane_f16(__s2_286, __p3_286)}); \
   __ret_286; \
 })
 #else
-#define vfmlsl_lane_low_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
-  float32x2_t __s0_287 = __p0_287; \
-  float16x4_t __s1_287 = __p1_287; \
-  float16x4_t __s2_287 = __p2_287; \
-  float32x2_t __rev0_287;  __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 1, 0); \
-  float16x4_t __rev1_287;  __rev1_287 = __builtin_shufflevector(__s1_287, __s1_287, 3, 2, 1, 0); \
-  float16x4_t __rev2_287;  __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 3, 2, 1, 0); \
-  float32x2_t __ret_287; \
-  __ret_287 = __noswap_vfmlsl_low_f16(__rev0_287, __rev1_287, (float16x4_t) {__noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287)}); \
-  __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 1, 0); \
+#define vfmlslq_laneq_low_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
+  float32x4_t __s0_287 = __p0_287; \
+  float16x8_t __s1_287 = __p1_287; \
+  float16x8_t __s2_287 = __p2_287; \
+  float32x4_t __rev0_287;  __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 3, 2, 1, 0); \
+  float16x8_t __rev1_287;  __rev1_287 = __builtin_shufflevector(__s1_287, __s1_287, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16x8_t __rev2_287;  __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x4_t __ret_287; \
+  __ret_287 = __noswap_vfmlslq_low_f16(__rev0_287, __rev1_287, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287), __noswap_vgetq_lane_f16(__rev2_287, __p3_287)}); \
+  __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 3, 2, 1, 0); \
   __ret_287; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_high_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
-  float32x4_t __s0_288 = __p0_288; \
-  float16x8_t __s1_288 = __p1_288; \
+#define vfmlsl_laneq_low_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
+  float32x2_t __s0_288 = __p0_288; \
+  float16x4_t __s1_288 = __p1_288; \
   float16x8_t __s2_288 = __p2_288; \
-  float32x4_t __ret_288; \
-  __ret_288 = vfmlslq_high_f16(__s0_288, __s1_288, (float16x8_t) {vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288)}); \
+  float32x2_t __ret_288; \
+  __ret_288 = vfmlsl_low_f16(__s0_288, __s1_288, (float16x4_t) {vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288)}); \
   __ret_288; \
 })
 #else
-#define vfmlslq_laneq_high_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
-  float32x4_t __s0_289 = __p0_289; \
-  float16x8_t __s1_289 = __p1_289; \
+#define vfmlsl_laneq_low_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
+  float32x2_t __s0_289 = __p0_289; \
+  float16x4_t __s1_289 = __p1_289; \
   float16x8_t __s2_289 = __p2_289; \
-  float32x4_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 3, 2, 1, 0); \
-  float16x8_t __rev1_289;  __rev1_289 = __builtin_shufflevector(__s1_289, __s1_289, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float32x2_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 1, 0); \
+  float16x4_t __rev1_289;  __rev1_289 = __builtin_shufflevector(__s1_289, __s1_289, 3, 2, 1, 0); \
   float16x8_t __rev2_289;  __rev2_289 = __builtin_shufflevector(__s2_289, __s2_289, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_289; \
-  __ret_289 = __noswap_vfmlslq_high_f16(__rev0_289, __rev1_289, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289)}); \
-  __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 3, 2, 1, 0); \
+  float32x2_t __ret_289; \
+  __ret_289 = __noswap_vfmlsl_low_f16(__rev0_289, __rev1_289, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289)}); \
+  __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 1, 0); \
   __ret_289; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_high_f16(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \
-  float32x2_t __s0_290 = __p0_290; \
-  float16x4_t __s1_290 = __p1_290; \
-  float16x8_t __s2_290 = __p2_290; \
-  float32x2_t __ret_290; \
-  __ret_290 = vfmlsl_high_f16(__s0_290, __s1_290, (float16x4_t) {vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290)}); \
-  __ret_290; \
-})
-#else
-#define vfmlsl_laneq_high_f16(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \
-  float32x2_t __s0_291 = __p0_291; \
-  float16x4_t __s1_291 = __p1_291; \
-  float16x8_t __s2_291 = __p2_291; \
-  float32x2_t __rev0_291;  __rev0_291 = __builtin_shufflevector(__s0_291, __s0_291, 1, 0); \
-  float16x4_t __rev1_291;  __rev1_291 = __builtin_shufflevector(__s1_291, __s1_291, 3, 2, 1, 0); \
-  float16x8_t __rev2_291;  __rev2_291 = __builtin_shufflevector(__s2_291, __s2_291, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_291; \
-  __ret_291 = __noswap_vfmlsl_high_f16(__rev0_291, __rev1_291, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291)}); \
-  __ret_291 = __builtin_shufflevector(__ret_291, __ret_291, 1, 0); \
-  __ret_291; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_low_f16(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \
-  float32x4_t __s0_292 = __p0_292; \
-  float16x8_t __s1_292 = __p1_292; \
-  float16x8_t __s2_292 = __p2_292; \
-  float32x4_t __ret_292; \
-  __ret_292 = vfmlslq_low_f16(__s0_292, __s1_292, (float16x8_t) {vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292)}); \
-  __ret_292; \
-})
-#else
-#define vfmlslq_laneq_low_f16(__p0_293, __p1_293, __p2_293, __p3_293) __extension__ ({ \
-  float32x4_t __s0_293 = __p0_293; \
-  float16x8_t __s1_293 = __p1_293; \
-  float16x8_t __s2_293 = __p2_293; \
-  float32x4_t __rev0_293;  __rev0_293 = __builtin_shufflevector(__s0_293, __s0_293, 3, 2, 1, 0); \
-  float16x8_t __rev1_293;  __rev1_293 = __builtin_shufflevector(__s1_293, __s1_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_293;  __rev2_293 = __builtin_shufflevector(__s2_293, __s2_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_293; \
-  __ret_293 = __noswap_vfmlslq_low_f16(__rev0_293, __rev1_293, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293)}); \
-  __ret_293 = __builtin_shufflevector(__ret_293, __ret_293, 3, 2, 1, 0); \
-  __ret_293; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_low_f16(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
-  float32x2_t __s0_294 = __p0_294; \
-  float16x4_t __s1_294 = __p1_294; \
-  float16x8_t __s2_294 = __p2_294; \
-  float32x2_t __ret_294; \
-  __ret_294 = vfmlsl_low_f16(__s0_294, __s1_294, (float16x4_t) {vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294)}); \
-  __ret_294; \
-})
-#else
-#define vfmlsl_laneq_low_f16(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
-  float32x2_t __s0_295 = __p0_295; \
-  float16x4_t __s1_295 = __p1_295; \
-  float16x8_t __s2_295 = __p2_295; \
-  float32x2_t __rev0_295;  __rev0_295 = __builtin_shufflevector(__s0_295, __s0_295, 1, 0); \
-  float16x4_t __rev1_295;  __rev1_295 = __builtin_shufflevector(__s1_295, __s1_295, 3, 2, 1, 0); \
-  float16x8_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_295; \
-  __ret_295 = __noswap_vfmlsl_low_f16(__rev0_295, __rev1_295, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295)}); \
-  __ret_295 = __builtin_shufflevector(__ret_295, __ret_295, 1, 0); \
-  __ret_295; \
-})
-#endif
-
 #endif
 #if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
-#define vmulh_lane_f16(__p0_296, __p1_296, __p2_296) __extension__ ({ \
-  float16_t __s0_296 = __p0_296; \
-  float16x4_t __s1_296 = __p1_296; \
-  float16_t __ret_296; \
-  __ret_296 = __s0_296 * vget_lane_f16(__s1_296, __p2_296); \
-  __ret_296; \
+#define vmulh_lane_f16(__p0_290, __p1_290, __p2_290) __extension__ ({ \
+  float16_t __s0_290 = __p0_290; \
+  float16x4_t __s1_290 = __p1_290; \
+  float16_t __ret_290; \
+  __ret_290 = __s0_290 * vget_lane_f16(__s1_290, __p2_290); \
+  __ret_290; \
 })
 #else
-#define vmulh_lane_f16(__p0_297, __p1_297, __p2_297) __extension__ ({ \
-  float16_t __s0_297 = __p0_297; \
-  float16x4_t __s1_297 = __p1_297; \
-  float16x4_t __rev1_297;  __rev1_297 = __builtin_shufflevector(__s1_297, __s1_297, 3, 2, 1, 0); \
-  float16_t __ret_297; \
-  __ret_297 = __s0_297 * __noswap_vget_lane_f16(__rev1_297, __p2_297); \
-  __ret_297; \
+#define vmulh_lane_f16(__p0_291, __p1_291, __p2_291) __extension__ ({ \
+  float16_t __s0_291 = __p0_291; \
+  float16x4_t __s1_291 = __p1_291; \
+  float16x4_t __rev1_291;  __rev1_291 = __builtin_shufflevector(__s1_291, __s1_291, 3, 2, 1, 0); \
+  float16_t __ret_291; \
+  __ret_291 = __s0_291 * __noswap_vget_lane_f16(__rev1_291, __p2_291); \
+  __ret_291; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vmulh_laneq_f16(__p0_298, __p1_298, __p2_298) __extension__ ({ \
-  float16_t __s0_298 = __p0_298; \
-  float16x8_t __s1_298 = __p1_298; \
-  float16_t __ret_298; \
-  __ret_298 = __s0_298 * vgetq_lane_f16(__s1_298, __p2_298); \
-  __ret_298; \
+#define vmulh_laneq_f16(__p0_292, __p1_292, __p2_292) __extension__ ({ \
+  float16_t __s0_292 = __p0_292; \
+  float16x8_t __s1_292 = __p1_292; \
+  float16_t __ret_292; \
+  __ret_292 = __s0_292 * vgetq_lane_f16(__s1_292, __p2_292); \
+  __ret_292; \
 })
 #else
-#define vmulh_laneq_f16(__p0_299, __p1_299, __p2_299) __extension__ ({ \
-  float16_t __s0_299 = __p0_299; \
-  float16x8_t __s1_299 = __p1_299; \
-  float16x8_t __rev1_299;  __rev1_299 = __builtin_shufflevector(__s1_299, __s1_299, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_299; \
-  __ret_299 = __s0_299 * __noswap_vgetq_lane_f16(__rev1_299, __p2_299); \
-  __ret_299; \
+#define vmulh_laneq_f16(__p0_293, __p1_293, __p2_293) __extension__ ({ \
+  float16_t __s0_293 = __p0_293; \
+  float16x8_t __s1_293 = __p1_293; \
+  float16x8_t __rev1_293;  __rev1_293 = __builtin_shufflevector(__s1_293, __s1_293, 7, 6, 5, 4, 3, 2, 1, 0); \
+  float16_t __ret_293; \
+  __ret_293 = __s0_293 * __noswap_vgetq_lane_f16(__rev1_293, __p2_293); \
+  __ret_293; \
 })
 #endif
 
 #endif
 #if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
   int32_t __ret;
   __ret = vqadds_s32(__p0, vqrdmulhs_s32(__p1, __p2));
   return __ret;
 }
-#else
-__ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqadds_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
   int16_t __ret;
   __ret = vqaddh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahs_lane_s32(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
+  int32_t __s0_294 = __p0_294; \
+  int32_t __s1_294 = __p1_294; \
+  int32x2_t __s2_294 = __p2_294; \
+  int32_t __ret_294; \
+  __ret_294 = vqadds_s32(__s0_294, vqrdmulhs_s32(__s1_294, vget_lane_s32(__s2_294, __p3_294))); \
+  __ret_294; \
+})
 #else
-__ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqaddh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
+#define vqrdmlahs_lane_s32(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
+  int32_t __s0_295 = __p0_295; \
+  int32_t __s1_295 = __p1_295; \
+  int32x2_t __s2_295 = __p2_295; \
+  int32x2_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 1, 0); \
+  int32_t __ret_295; \
+  __ret_295 = vqadds_s32(__s0_295, vqrdmulhs_s32(__s1_295, __noswap_vget_lane_s32(__rev2_295, __p3_295))); \
+  __ret_295; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_lane_s32(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
-  int32_t __s0_300 = __p0_300; \
-  int32_t __s1_300 = __p1_300; \
-  int32x2_t __s2_300 = __p2_300; \
-  int32_t __ret_300; \
-  __ret_300 = vqadds_s32(__s0_300, vqrdmulhs_s32(__s1_300, vget_lane_s32(__s2_300, __p3_300))); \
+#define vqrdmlahh_lane_s16(__p0_296, __p1_296, __p2_296, __p3_296) __extension__ ({ \
+  int16_t __s0_296 = __p0_296; \
+  int16_t __s1_296 = __p1_296; \
+  int16x4_t __s2_296 = __p2_296; \
+  int16_t __ret_296; \
+  __ret_296 = vqaddh_s16(__s0_296, vqrdmulhh_s16(__s1_296, vget_lane_s16(__s2_296, __p3_296))); \
+  __ret_296; \
+})
+#else
+#define vqrdmlahh_lane_s16(__p0_297, __p1_297, __p2_297, __p3_297) __extension__ ({ \
+  int16_t __s0_297 = __p0_297; \
+  int16_t __s1_297 = __p1_297; \
+  int16x4_t __s2_297 = __p2_297; \
+  int16x4_t __rev2_297;  __rev2_297 = __builtin_shufflevector(__s2_297, __s2_297, 3, 2, 1, 0); \
+  int16_t __ret_297; \
+  __ret_297 = vqaddh_s16(__s0_297, vqrdmulhh_s16(__s1_297, __noswap_vget_lane_s16(__rev2_297, __p3_297))); \
+  __ret_297; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahs_laneq_s32(__p0_298, __p1_298, __p2_298, __p3_298) __extension__ ({ \
+  int32_t __s0_298 = __p0_298; \
+  int32_t __s1_298 = __p1_298; \
+  int32x4_t __s2_298 = __p2_298; \
+  int32_t __ret_298; \
+  __ret_298 = vqadds_s32(__s0_298, vqrdmulhs_s32(__s1_298, vgetq_lane_s32(__s2_298, __p3_298))); \
+  __ret_298; \
+})
+#else
+#define vqrdmlahs_laneq_s32(__p0_299, __p1_299, __p2_299, __p3_299) __extension__ ({ \
+  int32_t __s0_299 = __p0_299; \
+  int32_t __s1_299 = __p1_299; \
+  int32x4_t __s2_299 = __p2_299; \
+  int32x4_t __rev2_299;  __rev2_299 = __builtin_shufflevector(__s2_299, __s2_299, 3, 2, 1, 0); \
+  int32_t __ret_299; \
+  __ret_299 = vqadds_s32(__s0_299, vqrdmulhs_s32(__s1_299, __noswap_vgetq_lane_s32(__rev2_299, __p3_299))); \
+  __ret_299; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlahh_laneq_s16(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
+  int16_t __s0_300 = __p0_300; \
+  int16_t __s1_300 = __p1_300; \
+  int16x8_t __s2_300 = __p2_300; \
+  int16_t __ret_300; \
+  __ret_300 = vqaddh_s16(__s0_300, vqrdmulhh_s16(__s1_300, vgetq_lane_s16(__s2_300, __p3_300))); \
   __ret_300; \
 })
 #else
-#define vqrdmlahs_lane_s32(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
-  int32_t __s0_301 = __p0_301; \
-  int32_t __s1_301 = __p1_301; \
-  int32x2_t __s2_301 = __p2_301; \
-  int32x2_t __rev2_301;  __rev2_301 = __builtin_shufflevector(__s2_301, __s2_301, 1, 0); \
-  int32_t __ret_301; \
-  __ret_301 = __noswap_vqadds_s32(__s0_301, __noswap_vqrdmulhs_s32(__s1_301, __noswap_vget_lane_s32(__rev2_301, __p3_301))); \
+#define vqrdmlahh_laneq_s16(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
+  int16_t __s0_301 = __p0_301; \
+  int16_t __s1_301 = __p1_301; \
+  int16x8_t __s2_301 = __p2_301; \
+  int16x8_t __rev2_301;  __rev2_301 = __builtin_shufflevector(__s2_301, __s2_301, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_301; \
+  __ret_301 = vqaddh_s16(__s0_301, vqrdmulhh_s16(__s1_301, __noswap_vgetq_lane_s16(__rev2_301, __p3_301))); \
   __ret_301; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_lane_s16(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
-  int16_t __s0_302 = __p0_302; \
-  int16_t __s1_302 = __p1_302; \
-  int16x4_t __s2_302 = __p2_302; \
-  int16_t __ret_302; \
-  __ret_302 = vqaddh_s16(__s0_302, vqrdmulhh_s16(__s1_302, vget_lane_s16(__s2_302, __p3_302))); \
-  __ret_302; \
-})
-#else
-#define vqrdmlahh_lane_s16(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
-  int16_t __s0_303 = __p0_303; \
-  int16_t __s1_303 = __p1_303; \
-  int16x4_t __s2_303 = __p2_303; \
-  int16x4_t __rev2_303;  __rev2_303 = __builtin_shufflevector(__s2_303, __s2_303, 3, 2, 1, 0); \
-  int16_t __ret_303; \
-  __ret_303 = __noswap_vqaddh_s16(__s0_303, __noswap_vqrdmulhh_s16(__s1_303, __noswap_vget_lane_s16(__rev2_303, __p3_303))); \
-  __ret_303; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_laneq_s32(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
-  int32_t __s0_304 = __p0_304; \
-  int32_t __s1_304 = __p1_304; \
-  int32x4_t __s2_304 = __p2_304; \
-  int32_t __ret_304; \
-  __ret_304 = vqadds_s32(__s0_304, vqrdmulhs_s32(__s1_304, vgetq_lane_s32(__s2_304, __p3_304))); \
-  __ret_304; \
-})
-#else
-#define vqrdmlahs_laneq_s32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
-  int32_t __s0_305 = __p0_305; \
-  int32_t __s1_305 = __p1_305; \
-  int32x4_t __s2_305 = __p2_305; \
-  int32x4_t __rev2_305;  __rev2_305 = __builtin_shufflevector(__s2_305, __s2_305, 3, 2, 1, 0); \
-  int32_t __ret_305; \
-  __ret_305 = __noswap_vqadds_s32(__s0_305, __noswap_vqrdmulhs_s32(__s1_305, __noswap_vgetq_lane_s32(__rev2_305, __p3_305))); \
-  __ret_305; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_laneq_s16(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
-  int16_t __s0_306 = __p0_306; \
-  int16_t __s1_306 = __p1_306; \
-  int16x8_t __s2_306 = __p2_306; \
-  int16_t __ret_306; \
-  __ret_306 = vqaddh_s16(__s0_306, vqrdmulhh_s16(__s1_306, vgetq_lane_s16(__s2_306, __p3_306))); \
-  __ret_306; \
-})
-#else
-#define vqrdmlahh_laneq_s16(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
-  int16_t __s0_307 = __p0_307; \
-  int16_t __s1_307 = __p1_307; \
-  int16x8_t __s2_307 = __p2_307; \
-  int16x8_t __rev2_307;  __rev2_307 = __builtin_shufflevector(__s2_307, __s2_307, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_307; \
-  __ret_307 = __noswap_vqaddh_s16(__s0_307, __noswap_vqrdmulhh_s16(__s1_307, __noswap_vgetq_lane_s16(__rev2_307, __p3_307))); \
-  __ret_307; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
   int32_t __ret;
   __ret = vqsubs_s32(__p0, vqrdmulhs_s32(__p1, __p2));
   return __ret;
 }
-#else
-__ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqsubs_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
   int16_t __ret;
   __ret = vqsubh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
   return __ret;
 }
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshs_lane_s32(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
+  int32_t __s0_302 = __p0_302; \
+  int32_t __s1_302 = __p1_302; \
+  int32x2_t __s2_302 = __p2_302; \
+  int32_t __ret_302; \
+  __ret_302 = vqsubs_s32(__s0_302, vqrdmulhs_s32(__s1_302, vget_lane_s32(__s2_302, __p3_302))); \
+  __ret_302; \
+})
 #else
-__ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqsubh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
+#define vqrdmlshs_lane_s32(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
+  int32_t __s0_303 = __p0_303; \
+  int32_t __s1_303 = __p1_303; \
+  int32x2_t __s2_303 = __p2_303; \
+  int32x2_t __rev2_303;  __rev2_303 = __builtin_shufflevector(__s2_303, __s2_303, 1, 0); \
+  int32_t __ret_303; \
+  __ret_303 = vqsubs_s32(__s0_303, vqrdmulhs_s32(__s1_303, __noswap_vget_lane_s32(__rev2_303, __p3_303))); \
+  __ret_303; \
+})
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_lane_s32(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
-  int32_t __s0_308 = __p0_308; \
-  int32_t __s1_308 = __p1_308; \
-  int32x2_t __s2_308 = __p2_308; \
-  int32_t __ret_308; \
-  __ret_308 = vqsubs_s32(__s0_308, vqrdmulhs_s32(__s1_308, vget_lane_s32(__s2_308, __p3_308))); \
+#define vqrdmlshh_lane_s16(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
+  int16_t __s0_304 = __p0_304; \
+  int16_t __s1_304 = __p1_304; \
+  int16x4_t __s2_304 = __p2_304; \
+  int16_t __ret_304; \
+  __ret_304 = vqsubh_s16(__s0_304, vqrdmulhh_s16(__s1_304, vget_lane_s16(__s2_304, __p3_304))); \
+  __ret_304; \
+})
+#else
+#define vqrdmlshh_lane_s16(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
+  int16_t __s0_305 = __p0_305; \
+  int16_t __s1_305 = __p1_305; \
+  int16x4_t __s2_305 = __p2_305; \
+  int16x4_t __rev2_305;  __rev2_305 = __builtin_shufflevector(__s2_305, __s2_305, 3, 2, 1, 0); \
+  int16_t __ret_305; \
+  __ret_305 = vqsubh_s16(__s0_305, vqrdmulhh_s16(__s1_305, __noswap_vget_lane_s16(__rev2_305, __p3_305))); \
+  __ret_305; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshs_laneq_s32(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
+  int32_t __s0_306 = __p0_306; \
+  int32_t __s1_306 = __p1_306; \
+  int32x4_t __s2_306 = __p2_306; \
+  int32_t __ret_306; \
+  __ret_306 = vqsubs_s32(__s0_306, vqrdmulhs_s32(__s1_306, vgetq_lane_s32(__s2_306, __p3_306))); \
+  __ret_306; \
+})
+#else
+#define vqrdmlshs_laneq_s32(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
+  int32_t __s0_307 = __p0_307; \
+  int32_t __s1_307 = __p1_307; \
+  int32x4_t __s2_307 = __p2_307; \
+  int32x4_t __rev2_307;  __rev2_307 = __builtin_shufflevector(__s2_307, __s2_307, 3, 2, 1, 0); \
+  int32_t __ret_307; \
+  __ret_307 = vqsubs_s32(__s0_307, vqrdmulhs_s32(__s1_307, __noswap_vgetq_lane_s32(__rev2_307, __p3_307))); \
+  __ret_307; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vqrdmlshh_laneq_s16(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
+  int16_t __s0_308 = __p0_308; \
+  int16_t __s1_308 = __p1_308; \
+  int16x8_t __s2_308 = __p2_308; \
+  int16_t __ret_308; \
+  __ret_308 = vqsubh_s16(__s0_308, vqrdmulhh_s16(__s1_308, vgetq_lane_s16(__s2_308, __p3_308))); \
   __ret_308; \
 })
 #else
-#define vqrdmlshs_lane_s32(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
-  int32_t __s0_309 = __p0_309; \
-  int32_t __s1_309 = __p1_309; \
-  int32x2_t __s2_309 = __p2_309; \
-  int32x2_t __rev2_309;  __rev2_309 = __builtin_shufflevector(__s2_309, __s2_309, 1, 0); \
-  int32_t __ret_309; \
-  __ret_309 = __noswap_vqsubs_s32(__s0_309, __noswap_vqrdmulhs_s32(__s1_309, __noswap_vget_lane_s32(__rev2_309, __p3_309))); \
+#define vqrdmlshh_laneq_s16(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
+  int16_t __s0_309 = __p0_309; \
+  int16_t __s1_309 = __p1_309; \
+  int16x8_t __s2_309 = __p2_309; \
+  int16x8_t __rev2_309;  __rev2_309 = __builtin_shufflevector(__s2_309, __s2_309, 7, 6, 5, 4, 3, 2, 1, 0); \
+  int16_t __ret_309; \
+  __ret_309 = vqsubh_s16(__s0_309, vqrdmulhh_s16(__s1_309, __noswap_vgetq_lane_s16(__rev2_309, __p3_309))); \
   __ret_309; \
 })
 #endif
 
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_lane_s16(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
-  int16_t __s0_310 = __p0_310; \
-  int16_t __s1_310 = __p1_310; \
-  int16x4_t __s2_310 = __p2_310; \
-  int16_t __ret_310; \
-  __ret_310 = vqsubh_s16(__s0_310, vqrdmulhh_s16(__s1_310, vget_lane_s16(__s2_310, __p3_310))); \
-  __ret_310; \
-})
-#else
-#define vqrdmlshh_lane_s16(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
-  int16_t __s0_311 = __p0_311; \
-  int16_t __s1_311 = __p1_311; \
-  int16x4_t __s2_311 = __p2_311; \
-  int16x4_t __rev2_311;  __rev2_311 = __builtin_shufflevector(__s2_311, __s2_311, 3, 2, 1, 0); \
-  int16_t __ret_311; \
-  __ret_311 = __noswap_vqsubh_s16(__s0_311, __noswap_vqrdmulhh_s16(__s1_311, __noswap_vget_lane_s16(__rev2_311, __p3_311))); \
-  __ret_311; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_laneq_s32(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
-  int32_t __s0_312 = __p0_312; \
-  int32_t __s1_312 = __p1_312; \
-  int32x4_t __s2_312 = __p2_312; \
-  int32_t __ret_312; \
-  __ret_312 = vqsubs_s32(__s0_312, vqrdmulhs_s32(__s1_312, vgetq_lane_s32(__s2_312, __p3_312))); \
-  __ret_312; \
-})
-#else
-#define vqrdmlshs_laneq_s32(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
-  int32_t __s0_313 = __p0_313; \
-  int32_t __s1_313 = __p1_313; \
-  int32x4_t __s2_313 = __p2_313; \
-  int32x4_t __rev2_313;  __rev2_313 = __builtin_shufflevector(__s2_313, __s2_313, 3, 2, 1, 0); \
-  int32_t __ret_313; \
-  __ret_313 = __noswap_vqsubs_s32(__s0_313, __noswap_vqrdmulhs_s32(__s1_313, __noswap_vgetq_lane_s32(__rev2_313, __p3_313))); \
-  __ret_313; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_laneq_s16(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
-  int16_t __s0_314 = __p0_314; \
-  int16_t __s1_314 = __p1_314; \
-  int16x8_t __s2_314 = __p2_314; \
-  int16_t __ret_314; \
-  __ret_314 = vqsubh_s16(__s0_314, vqrdmulhh_s16(__s1_314, vgetq_lane_s16(__s2_314, __p3_314))); \
-  __ret_314; \
-})
-#else
-#define vqrdmlshh_laneq_s16(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
-  int16_t __s0_315 = __p0_315; \
-  int16_t __s1_315 = __p1_315; \
-  int16x8_t __s2_315 = __p2_315; \
-  int16x8_t __rev2_315;  __rev2_315 = __builtin_shufflevector(__s2_315, __s2_315, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_315; \
-  __ret_315 = __noswap_vqsubh_s16(__s0_315, __noswap_vqrdmulhh_s16(__s1_315, __noswap_vgetq_lane_s16(__rev2_315, __p3_315))); \
-  __ret_315; \
-})
-#endif
-
 #endif
 #if defined(__aarch64__)
 #ifdef __LITTLE_ENDIAN__
@@ -72870,162 +62089,140 @@
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_p64(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
+#define vcopyq_lane_p64(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
+  poly64x2_t __s0_310 = __p0_310; \
+  poly64x1_t __s2_310 = __p2_310; \
+  poly64x2_t __ret_310; \
+  __ret_310 = vsetq_lane_p64(vget_lane_p64(__s2_310, __p3_310), __s0_310, __p1_310); \
+  __ret_310; \
+})
+#else
+#define vcopyq_lane_p64(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
+  poly64x2_t __s0_311 = __p0_311; \
+  poly64x1_t __s2_311 = __p2_311; \
+  poly64x2_t __rev0_311;  __rev0_311 = __builtin_shufflevector(__s0_311, __s0_311, 1, 0); \
+  poly64x2_t __ret_311; \
+  __ret_311 = __noswap_vsetq_lane_p64(vget_lane_p64(__s2_311, __p3_311), __rev0_311, __p1_311); \
+  __ret_311 = __builtin_shufflevector(__ret_311, __ret_311, 1, 0); \
+  __ret_311; \
+})
+#endif
+
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_lane_f64(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
+  float64x2_t __s0_312 = __p0_312; \
+  float64x1_t __s2_312 = __p2_312; \
+  float64x2_t __ret_312; \
+  __ret_312 = vsetq_lane_f64(vget_lane_f64(__s2_312, __p3_312), __s0_312, __p1_312); \
+  __ret_312; \
+})
+#else
+#define vcopyq_lane_f64(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
+  float64x2_t __s0_313 = __p0_313; \
+  float64x1_t __s2_313 = __p2_313; \
+  float64x2_t __rev0_313;  __rev0_313 = __builtin_shufflevector(__s0_313, __s0_313, 1, 0); \
+  float64x2_t __ret_313; \
+  __ret_313 = __noswap_vsetq_lane_f64(vget_lane_f64(__s2_313, __p3_313), __rev0_313, __p1_313); \
+  __ret_313 = __builtin_shufflevector(__ret_313, __ret_313, 1, 0); \
+  __ret_313; \
+})
+#endif
+
+#define vcopy_lane_p64(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
+  poly64x1_t __s0_314 = __p0_314; \
+  poly64x1_t __s2_314 = __p2_314; \
+  poly64x1_t __ret_314; \
+  __ret_314 = vset_lane_p64(vget_lane_p64(__s2_314, __p3_314), __s0_314, __p1_314); \
+  __ret_314; \
+})
+#define vcopy_lane_f64(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
+  float64x1_t __s0_315 = __p0_315; \
+  float64x1_t __s2_315 = __p2_315; \
+  float64x1_t __ret_315; \
+  __ret_315 = vset_lane_f64(vget_lane_f64(__s2_315, __p3_315), __s0_315, __p1_315); \
+  __ret_315; \
+})
+#ifdef __LITTLE_ENDIAN__
+#define vcopyq_laneq_p64(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
   poly64x2_t __s0_316 = __p0_316; \
-  poly64x1_t __s2_316 = __p2_316; \
+  poly64x2_t __s2_316 = __p2_316; \
   poly64x2_t __ret_316; \
-  __ret_316 = vsetq_lane_p64(vget_lane_p64(__s2_316, __p3_316), __s0_316, __p1_316); \
+  __ret_316 = vsetq_lane_p64(vgetq_lane_p64(__s2_316, __p3_316), __s0_316, __p1_316); \
   __ret_316; \
 })
 #else
-#define vcopyq_lane_p64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
+#define vcopyq_laneq_p64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
   poly64x2_t __s0_317 = __p0_317; \
-  poly64x1_t __s2_317 = __p2_317; \
+  poly64x2_t __s2_317 = __p2_317; \
   poly64x2_t __rev0_317;  __rev0_317 = __builtin_shufflevector(__s0_317, __s0_317, 1, 0); \
+  poly64x2_t __rev2_317;  __rev2_317 = __builtin_shufflevector(__s2_317, __s2_317, 1, 0); \
   poly64x2_t __ret_317; \
-  __ret_317 = __noswap_vsetq_lane_p64(__noswap_vget_lane_p64(__s2_317, __p3_317), __rev0_317, __p1_317); \
+  __ret_317 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_317, __p3_317), __rev0_317, __p1_317); \
   __ret_317 = __builtin_shufflevector(__ret_317, __ret_317, 1, 0); \
   __ret_317; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_f64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
+#define vcopyq_laneq_f64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
   float64x2_t __s0_318 = __p0_318; \
-  float64x1_t __s2_318 = __p2_318; \
+  float64x2_t __s2_318 = __p2_318; \
   float64x2_t __ret_318; \
-  __ret_318 = vsetq_lane_f64(vget_lane_f64(__s2_318, __p3_318), __s0_318, __p1_318); \
+  __ret_318 = vsetq_lane_f64(vgetq_lane_f64(__s2_318, __p3_318), __s0_318, __p1_318); \
   __ret_318; \
 })
 #else
-#define vcopyq_lane_f64(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
+#define vcopyq_laneq_f64(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
   float64x2_t __s0_319 = __p0_319; \
-  float64x1_t __s2_319 = __p2_319; \
+  float64x2_t __s2_319 = __p2_319; \
   float64x2_t __rev0_319;  __rev0_319 = __builtin_shufflevector(__s0_319, __s0_319, 1, 0); \
+  float64x2_t __rev2_319;  __rev2_319 = __builtin_shufflevector(__s2_319, __s2_319, 1, 0); \
   float64x2_t __ret_319; \
-  __ret_319 = __noswap_vsetq_lane_f64(__noswap_vget_lane_f64(__s2_319, __p3_319), __rev0_319, __p1_319); \
+  __ret_319 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_319, __p3_319), __rev0_319, __p1_319); \
   __ret_319 = __builtin_shufflevector(__ret_319, __ret_319, 1, 0); \
   __ret_319; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_p64(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
+#define vcopy_laneq_p64(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
   poly64x1_t __s0_320 = __p0_320; \
-  poly64x1_t __s2_320 = __p2_320; \
+  poly64x2_t __s2_320 = __p2_320; \
   poly64x1_t __ret_320; \
-  __ret_320 = vset_lane_p64(vget_lane_p64(__s2_320, __p3_320), __s0_320, __p1_320); \
+  __ret_320 = vset_lane_p64(vgetq_lane_p64(__s2_320, __p3_320), __s0_320, __p1_320); \
   __ret_320; \
 })
 #else
-#define vcopy_lane_p64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
+#define vcopy_laneq_p64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
   poly64x1_t __s0_321 = __p0_321; \
-  poly64x1_t __s2_321 = __p2_321; \
+  poly64x2_t __s2_321 = __p2_321; \
+  poly64x2_t __rev2_321;  __rev2_321 = __builtin_shufflevector(__s2_321, __s2_321, 1, 0); \
   poly64x1_t __ret_321; \
-  __ret_321 = __noswap_vset_lane_p64(__noswap_vget_lane_p64(__s2_321, __p3_321), __s0_321, __p1_321); \
+  __ret_321 = vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_321, __p3_321), __s0_321, __p1_321); \
   __ret_321; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_f64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
+#define vcopy_laneq_f64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
   float64x1_t __s0_322 = __p0_322; \
-  float64x1_t __s2_322 = __p2_322; \
+  float64x2_t __s2_322 = __p2_322; \
   float64x1_t __ret_322; \
-  __ret_322 = vset_lane_f64(vget_lane_f64(__s2_322, __p3_322), __s0_322, __p1_322); \
+  __ret_322 = vset_lane_f64(vgetq_lane_f64(__s2_322, __p3_322), __s0_322, __p1_322); \
   __ret_322; \
 })
 #else
-#define vcopy_lane_f64(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
+#define vcopy_laneq_f64(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
   float64x1_t __s0_323 = __p0_323; \
-  float64x1_t __s2_323 = __p2_323; \
+  float64x2_t __s2_323 = __p2_323; \
+  float64x2_t __rev2_323;  __rev2_323 = __builtin_shufflevector(__s2_323, __s2_323, 1, 0); \
   float64x1_t __ret_323; \
-  __ret_323 = __noswap_vset_lane_f64(__noswap_vget_lane_f64(__s2_323, __p3_323), __s0_323, __p1_323); \
+  __ret_323 = vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_323, __p3_323), __s0_323, __p1_323); \
   __ret_323; \
 })
 #endif
 
 #ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p64(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \
-  poly64x2_t __s0_324 = __p0_324; \
-  poly64x2_t __s2_324 = __p2_324; \
-  poly64x2_t __ret_324; \
-  __ret_324 = vsetq_lane_p64(vgetq_lane_p64(__s2_324, __p3_324), __s0_324, __p1_324); \
-  __ret_324; \
-})
-#else
-#define vcopyq_laneq_p64(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \
-  poly64x2_t __s0_325 = __p0_325; \
-  poly64x2_t __s2_325 = __p2_325; \
-  poly64x2_t __rev0_325;  __rev0_325 = __builtin_shufflevector(__s0_325, __s0_325, 1, 0); \
-  poly64x2_t __rev2_325;  __rev2_325 = __builtin_shufflevector(__s2_325, __s2_325, 1, 0); \
-  poly64x2_t __ret_325; \
-  __ret_325 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_325, __p3_325), __rev0_325, __p1_325); \
-  __ret_325 = __builtin_shufflevector(__ret_325, __ret_325, 1, 0); \
-  __ret_325; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f64(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \
-  float64x2_t __s0_326 = __p0_326; \
-  float64x2_t __s2_326 = __p2_326; \
-  float64x2_t __ret_326; \
-  __ret_326 = vsetq_lane_f64(vgetq_lane_f64(__s2_326, __p3_326), __s0_326, __p1_326); \
-  __ret_326; \
-})
-#else
-#define vcopyq_laneq_f64(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \
-  float64x2_t __s0_327 = __p0_327; \
-  float64x2_t __s2_327 = __p2_327; \
-  float64x2_t __rev0_327;  __rev0_327 = __builtin_shufflevector(__s0_327, __s0_327, 1, 0); \
-  float64x2_t __rev2_327;  __rev2_327 = __builtin_shufflevector(__s2_327, __s2_327, 1, 0); \
-  float64x2_t __ret_327; \
-  __ret_327 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_327, __p3_327), __rev0_327, __p1_327); \
-  __ret_327 = __builtin_shufflevector(__ret_327, __ret_327, 1, 0); \
-  __ret_327; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p64(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \
-  poly64x1_t __s0_328 = __p0_328; \
-  poly64x2_t __s2_328 = __p2_328; \
-  poly64x1_t __ret_328; \
-  __ret_328 = vset_lane_p64(vgetq_lane_p64(__s2_328, __p3_328), __s0_328, __p1_328); \
-  __ret_328; \
-})
-#else
-#define vcopy_laneq_p64(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \
-  poly64x1_t __s0_329 = __p0_329; \
-  poly64x2_t __s2_329 = __p2_329; \
-  poly64x2_t __rev2_329;  __rev2_329 = __builtin_shufflevector(__s2_329, __s2_329, 1, 0); \
-  poly64x1_t __ret_329; \
-  __ret_329 = __noswap_vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_329, __p3_329), __s0_329, __p1_329); \
-  __ret_329; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f64(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \
-  float64x1_t __s0_330 = __p0_330; \
-  float64x2_t __s2_330 = __p2_330; \
-  float64x1_t __ret_330; \
-  __ret_330 = vset_lane_f64(vgetq_lane_f64(__s2_330, __p3_330), __s0_330, __p1_330); \
-  __ret_330; \
-})
-#else
-#define vcopy_laneq_f64(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \
-  float64x1_t __s0_331 = __p0_331; \
-  float64x2_t __s2_331 = __p2_331; \
-  float64x2_t __rev2_331;  __rev2_331 = __builtin_shufflevector(__s2_331, __s2_331, 1, 0); \
-  float64x1_t __ret_331; \
-  __ret_331 = __noswap_vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_331, __p3_331), __s0_331, __p1_331); \
-  __ret_331; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
 __ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
   uint16x8_t __ret;
   __ret = vmlal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
@@ -73377,52 +62574,38 @@
 }
 #endif
 
+#define vmulx_lane_f64(__p0_324, __p1_324, __p2_324) __extension__ ({ \
+  float64x1_t __s0_324 = __p0_324; \
+  float64x1_t __s1_324 = __p1_324; \
+  float64x1_t __ret_324; \
+  float64_t __x_324 = vget_lane_f64(__s0_324, 0); \
+  float64_t __y_324 = vget_lane_f64(__s1_324, __p2_324); \
+  float64_t __z_324 = vmulxd_f64(__x_324, __y_324); \
+  __ret_324 = vset_lane_f64(__z_324, __s0_324, __p2_324); \
+  __ret_324; \
+})
 #ifdef __LITTLE_ENDIAN__
-#define vmulx_lane_f64(__p0_332, __p1_332, __p2_332) __extension__ ({ \
-  float64x1_t __s0_332 = __p0_332; \
-  float64x1_t __s1_332 = __p1_332; \
-  float64x1_t __ret_332; \
-  float64_t __x_332 = vget_lane_f64(__s0_332, 0); \
-  float64_t __y_332 = vget_lane_f64(__s1_332, __p2_332); \
-  float64_t __z_332 = vmulxd_f64(__x_332, __y_332); \
-  __ret_332 = vset_lane_f64(__z_332, __s0_332, __p2_332); \
-  __ret_332; \
+#define vmulx_laneq_f64(__p0_325, __p1_325, __p2_325) __extension__ ({ \
+  float64x1_t __s0_325 = __p0_325; \
+  float64x2_t __s1_325 = __p1_325; \
+  float64x1_t __ret_325; \
+  float64_t __x_325 = vget_lane_f64(__s0_325, 0); \
+  float64_t __y_325 = vgetq_lane_f64(__s1_325, __p2_325); \
+  float64_t __z_325 = vmulxd_f64(__x_325, __y_325); \
+  __ret_325 = vset_lane_f64(__z_325, __s0_325, 0); \
+  __ret_325; \
 })
 #else
-#define vmulx_lane_f64(__p0_333, __p1_333, __p2_333) __extension__ ({ \
-  float64x1_t __s0_333 = __p0_333; \
-  float64x1_t __s1_333 = __p1_333; \
-  float64x1_t __ret_333; \
-  float64_t __x_333 = __noswap_vget_lane_f64(__s0_333, 0); \
-  float64_t __y_333 = __noswap_vget_lane_f64(__s1_333, __p2_333); \
-  float64_t __z_333 = __noswap_vmulxd_f64(__x_333, __y_333); \
-  __ret_333 = __noswap_vset_lane_f64(__z_333, __s0_333, __p2_333); \
-  __ret_333; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_laneq_f64(__p0_334, __p1_334, __p2_334) __extension__ ({ \
-  float64x1_t __s0_334 = __p0_334; \
-  float64x2_t __s1_334 = __p1_334; \
-  float64x1_t __ret_334; \
-  float64_t __x_334 = vget_lane_f64(__s0_334, 0); \
-  float64_t __y_334 = vgetq_lane_f64(__s1_334, __p2_334); \
-  float64_t __z_334 = vmulxd_f64(__x_334, __y_334); \
-  __ret_334 = vset_lane_f64(__z_334, __s0_334, 0); \
-  __ret_334; \
-})
-#else
-#define vmulx_laneq_f64(__p0_335, __p1_335, __p2_335) __extension__ ({ \
-  float64x1_t __s0_335 = __p0_335; \
-  float64x2_t __s1_335 = __p1_335; \
-  float64x2_t __rev1_335;  __rev1_335 = __builtin_shufflevector(__s1_335, __s1_335, 1, 0); \
-  float64x1_t __ret_335; \
-  float64_t __x_335 = __noswap_vget_lane_f64(__s0_335, 0); \
-  float64_t __y_335 = __noswap_vgetq_lane_f64(__rev1_335, __p2_335); \
-  float64_t __z_335 = __noswap_vmulxd_f64(__x_335, __y_335); \
-  __ret_335 = __noswap_vset_lane_f64(__z_335, __s0_335, 0); \
-  __ret_335; \
+#define vmulx_laneq_f64(__p0_326, __p1_326, __p2_326) __extension__ ({ \
+  float64x1_t __s0_326 = __p0_326; \
+  float64x2_t __s1_326 = __p1_326; \
+  float64x2_t __rev1_326;  __rev1_326 = __builtin_shufflevector(__s1_326, __s1_326, 1, 0); \
+  float64x1_t __ret_326; \
+  float64_t __x_326 = vget_lane_f64(__s0_326, 0); \
+  float64_t __y_326 = __noswap_vgetq_lane_f64(__rev1_326, __p2_326); \
+  float64_t __z_326 = vmulxd_f64(__x_326, __y_326); \
+  __ret_326 = vset_lane_f64(__z_326, __s0_326, 0); \
+  __ret_326; \
 })
 #endif
 
diff --git a/linux-x86/lib64/clang/9.0.7/include/armintr.h b/linux-x86/lib64/clang/10.0.1/include/armintr.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/armintr.h
rename to linux-x86/lib64/clang/10.0.1/include/armintr.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx2intrin.h b/linux-x86/lib64/clang/10.0.1/include/avx2intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx2intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx2intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512bf16intrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512bf16intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512bf16intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512bf16intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512bitalgintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512bitalgintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512bitalgintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512bitalgintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512bwintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512bwintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512bwintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512bwintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512cdintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512cdintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512cdintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512cdintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512dqintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512dqintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512dqintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512dqintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512erintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512erintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512erintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512erintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512fintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/avx512fintrin.h
index 132761f..4e341a1 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/avx512fintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/avx512fintrin.h
@@ -8436,7 +8436,7 @@
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_si512 (__m512i * __P, __m512i __A)
+_mm512_stream_si512 (void * __P, __m512i __A)
 {
   typedef __v8di __v8di_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v8di_aligned)__A, (__v8di_aligned*)__P);
@@ -8450,14 +8450,14 @@
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_pd (double *__P, __m512d __A)
+_mm512_stream_pd (void *__P, __m512d __A)
 {
   typedef __v8df __v8df_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v8df_aligned)__A, (__v8df_aligned*)__P);
 }
 
 static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_ps (float *__P, __m512 __A)
+_mm512_stream_ps (void *__P, __m512 __A)
 {
   typedef __v16sf __v16sf_aligned __attribute__((aligned(64)));
   __builtin_nontemporal_store((__v16sf_aligned)__A, (__v16sf_aligned*)__P);
@@ -9659,6 +9659,23 @@
 }
 #undef _mm512_mask_reduce_operator
 
+/// Moves the least significant 32 bits of a vector of [16 x i32] to a
+///    32-bit signed integer value.
+///
+/// \headerfile <x86intrin.h>
+///
+/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
+///
+/// \param __A
+///    A vector of [16 x i32]. The least significant 32 bits are moved to the
+///    destination.
+/// \returns A 32-bit signed integer containing the moved value.
+static __inline__ int __DEFAULT_FN_ATTRS512
+_mm512_cvtsi512_si32(__m512i __A) {
+  __v16si __b = (__v16si)__A;
+  return __b[0];
+}
+
 #undef __DEFAULT_FN_ATTRS512
 #undef __DEFAULT_FN_ATTRS128
 #undef __DEFAULT_FN_ATTRS
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512ifmaintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512ifmaintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512ifmaintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512ifmaintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512ifmavlintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512ifmavlintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512ifmavlintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512ifmavlintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512pfintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512pfintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512pfintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512pfintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vbmi2intrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vbmi2intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vbmi2intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vbmi2intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vbmiintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vbmiintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vbmiintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vbmiintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vbmivlintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vbmivlintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vbmivlintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vbmivlintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlbf16intrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlbf16intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlbf16intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlbf16intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlbitalgintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlbitalgintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlbitalgintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlbitalgintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlbwintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlbwintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlbwintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlbwintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlcdintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlcdintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlcdintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlcdintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vldqintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vldqintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vldqintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vldqintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlvbmi2intrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlvbmi2intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlvbmi2intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlvbmi2intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlvnniintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlvnniintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlvnniintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlvnniintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vlvp2intersectintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vlvp2intersectintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vlvp2intersectintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vlvp2intersectintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vnniintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vnniintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vnniintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vnniintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vp2intersectintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vp2intersectintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vp2intersectintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vp2intersectintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vpopcntdqintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vpopcntdqintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vpopcntdqintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vpopcntdqintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512vpopcntdqvlintrin.h b/linux-x86/lib64/clang/10.0.1/include/avx512vpopcntdqvlintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avx512vpopcntdqvlintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avx512vpopcntdqvlintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/avxintrin.h b/linux-x86/lib64/clang/10.0.1/include/avxintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/avxintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/avxintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/bits/stdatomic.h b/linux-x86/lib64/clang/10.0.1/include/bits/stdatomic.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/bits/stdatomic.h
rename to linux-x86/lib64/clang/10.0.1/include/bits/stdatomic.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/bmi2intrin.h b/linux-x86/lib64/clang/10.0.1/include/bmi2intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/bmi2intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/bmi2intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/bmiintrin.h b/linux-x86/lib64/clang/10.0.1/include/bmiintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/bmiintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/bmiintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/cetintrin.h b/linux-x86/lib64/clang/10.0.1/include/cetintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/cetintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/cetintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/cldemoteintrin.h b/linux-x86/lib64/clang/10.0.1/include/cldemoteintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/cldemoteintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/cldemoteintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/clflushoptintrin.h b/linux-x86/lib64/clang/10.0.1/include/clflushoptintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/clflushoptintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/clflushoptintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/clwbintrin.h b/linux-x86/lib64/clang/10.0.1/include/clwbintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/clwbintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/clwbintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/clzerointrin.h b/linux-x86/lib64/clang/10.0.1/include/clzerointrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/clzerointrin.h
rename to linux-x86/lib64/clang/10.0.1/include/clzerointrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/cpuid.h b/linux-x86/lib64/clang/10.0.1/include/cpuid.h
similarity index 98%
copy from darwin-x86/lib64/clang/9.0.7/include/cpuid.h
copy to linux-x86/lib64/clang/10.0.1/include/cpuid.h
index 02ffac2..4ddd648 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/cpuid.h
+++ b/linux-x86/lib64/clang/10.0.1/include/cpuid.h
@@ -38,8 +38,8 @@
 #define signature_TM2_ecx 0x3638784d
 /* NSC:     "Geode by NSC" */
 #define signature_NSC_ebx 0x646f6547
-#define signature_NSC_edx 0x43534e20
-#define signature_NSC_ecx 0x79622065
+#define signature_NSC_edx 0x79622065
+#define signature_NSC_ecx 0x43534e20
 /* NEXGEN:  "NexGenDriven" */
 #define signature_NEXGEN_ebx 0x4778654e
 #define signature_NEXGEN_edx 0x72446e65
diff --git a/linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/algorithm b/linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/algorithm
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/algorithm
rename to linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/algorithm
diff --git a/linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/complex b/linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/complex
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/complex
rename to linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/complex
diff --git a/linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/new b/linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/new
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/cuda_wrappers/new
rename to linux-x86/lib64/clang/10.0.1/include/cuda_wrappers/new
diff --git a/darwin-x86/lib64/clang/9.0.7/include/emmintrin.h b/linux-x86/lib64/clang/10.0.1/include/emmintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/emmintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/emmintrin.h
index 3d55f5f..c8fefdf 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/emmintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/emmintrin.h
@@ -4029,7 +4029,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si64(void const *__p, __m128i __b)
+_mm_storeu_si64(void *__p, __m128i __b)
 {
   struct __storeu_si64 {
     long long __v;
@@ -4050,7 +4050,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si32(void const *__p, __m128i __b)
+_mm_storeu_si32(void *__p, __m128i __b)
 {
   struct __storeu_si32 {
     int __v;
@@ -4071,7 +4071,7 @@
 /// \param __b
 ///    A 128-bit integer vector containing the value to be stored.
 static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si16(void const *__p, __m128i __b)
+_mm_storeu_si16(void *__p, __m128i __b)
 {
   struct __storeu_si16 {
     short __v;
diff --git a/linux-x86/lib64/clang/9.0.7/include/enqcmdintrin.h b/linux-x86/lib64/clang/10.0.1/include/enqcmdintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/enqcmdintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/enqcmdintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/f16cintrin.h b/linux-x86/lib64/clang/10.0.1/include/f16cintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/f16cintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/f16cintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/float.h b/linux-x86/lib64/clang/10.0.1/include/float.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/float.h
rename to linux-x86/lib64/clang/10.0.1/include/float.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/fma4intrin.h b/linux-x86/lib64/clang/10.0.1/include/fma4intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/fma4intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/fma4intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/fmaintrin.h b/linux-x86/lib64/clang/10.0.1/include/fmaintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/fmaintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/fmaintrin.h
diff --git a/linux-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h b/linux-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h
new file mode 100644
index 0000000..59ef284
--- /dev/null
+++ b/linux-x86/lib64/clang/10.0.1/include/fuzzer/FuzzedDataProvider.h
@@ -0,0 +1,299 @@
+//===- FuzzedDataProvider.h - Utility header for fuzz targets ---*- C++ -* ===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+// A single header library providing an utility class to break up an array of
+// bytes. Whenever run on the same input, provides the same output, as long as
+// its methods are called in the same order, with the same arguments.
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
+#define LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
+
+#include <algorithm>
+#include <climits>
+#include <cstddef>
+#include <cstdint>
+#include <cstring>
+#include <initializer_list>
+#include <string>
+#include <type_traits>
+#include <utility>
+#include <vector>
+
+// In addition to the comments below, the API is also briefly documented at
+// https://github.com/google/fuzzing/blob/master/docs/split-inputs.md#fuzzed-data-provider
+class FuzzedDataProvider {
+ public:
+  // |data| is an array of length |size| that the FuzzedDataProvider wraps to
+  // provide more granular access. |data| must outlive the FuzzedDataProvider.
+  FuzzedDataProvider(const uint8_t *data, size_t size)
+      : data_ptr_(data), remaining_bytes_(size) {}
+  ~FuzzedDataProvider() = default;
+
+  // Returns a std::vector containing |num_bytes| of input data. If fewer than
+  // |num_bytes| of data remain, returns a shorter std::vector containing all
+  // of the data that's left. Can be used with any byte sized type, such as
+  // char, unsigned char, uint8_t, etc.
+  template <typename T> std::vector<T> ConsumeBytes(size_t num_bytes) {
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    return ConsumeBytes<T>(num_bytes, num_bytes);
+  }
+
+  // Similar to |ConsumeBytes|, but also appends the terminator value at the end
+  // of the resulting vector. Useful, when a mutable null-terminated C-string is
+  // needed, for example. But that is a rare case. Better avoid it, if possible,
+  // and prefer using |ConsumeBytes| or |ConsumeBytesAsString| methods.
+  template <typename T>
+  std::vector<T> ConsumeBytesWithTerminator(size_t num_bytes,
+                                            T terminator = 0) {
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    std::vector<T> result = ConsumeBytes<T>(num_bytes + 1, num_bytes);
+    result.back() = terminator;
+    return result;
+  }
+
+  // Returns a std::string containing |num_bytes| of input data. Using this and
+  // |.c_str()| on the resulting string is the best way to get an immutable
+  // null-terminated C string. If fewer than |num_bytes| of data remain, returns
+  // a shorter std::string containing all of the data that's left.
+  std::string ConsumeBytesAsString(size_t num_bytes) {
+    static_assert(sizeof(std::string::value_type) == sizeof(uint8_t),
+                  "ConsumeBytesAsString cannot convert the data to a string.");
+
+    num_bytes = std::min(num_bytes, remaining_bytes_);
+    std::string result(
+        reinterpret_cast<const std::string::value_type *>(data_ptr_),
+        num_bytes);
+    Advance(num_bytes);
+    return result;
+  }
+
+  // Returns a number in the range [min, max] by consuming bytes from the
+  // input data. The value might not be uniformly distributed in the given
+  // range. If there's no input data left, always returns |min|. |min| must
+  // be less than or equal to |max|.
+  template <typename T> T ConsumeIntegralInRange(T min, T max) {
+    static_assert(std::is_integral<T>::value, "An integral type is required.");
+    static_assert(sizeof(T) <= sizeof(uint64_t), "Unsupported integral type.");
+
+    if (min > max)
+      abort();
+
+    // Use the biggest type possible to hold the range and the result.
+    uint64_t range = static_cast<uint64_t>(max) - min;
+    uint64_t result = 0;
+    size_t offset = 0;
+
+    while (offset < sizeof(T) * CHAR_BIT && (range >> offset) > 0 &&
+           remaining_bytes_ != 0) {
+      // Pull bytes off the end of the seed data. Experimentally, this seems to
+      // allow the fuzzer to more easily explore the input space. This makes
+      // sense, since it works by modifying inputs that caused new code to run,
+      // and this data is often used to encode length of data read by
+      // |ConsumeBytes|. Separating out read lengths makes it easier modify the
+      // contents of the data that is actually read.
+      --remaining_bytes_;
+      result = (result << CHAR_BIT) | data_ptr_[remaining_bytes_];
+      offset += CHAR_BIT;
+    }
+
+    // Avoid division by 0, in case |range + 1| results in overflow.
+    if (range != std::numeric_limits<decltype(range)>::max())
+      result = result % (range + 1);
+
+    return static_cast<T>(min + result);
+  }
+
+  // Returns a std::string of length from 0 to |max_length|. When it runs out of
+  // input data, returns what remains of the input. Designed to be more stable
+  // with respect to a fuzzer inserting characters than just picking a random
+  // length and then consuming that many bytes with |ConsumeBytes|.
+  std::string ConsumeRandomLengthString(size_t max_length) {
+    // Reads bytes from the start of |data_ptr_|. Maps "\\" to "\", and maps "\"
+    // followed by anything else to the end of the string. As a result of this
+    // logic, a fuzzer can insert characters into the string, and the string
+    // will be lengthened to include those new characters, resulting in a more
+    // stable fuzzer than picking the length of a string independently from
+    // picking its contents.
+    std::string result;
+
+    // Reserve the anticipated capaticity to prevent several reallocations.
+    result.reserve(std::min(max_length, remaining_bytes_));
+    for (size_t i = 0; i < max_length && remaining_bytes_ != 0; ++i) {
+      char next = ConvertUnsignedToSigned<char>(data_ptr_[0]);
+      Advance(1);
+      if (next == '\\' && remaining_bytes_ != 0) {
+        next = ConvertUnsignedToSigned<char>(data_ptr_[0]);
+        Advance(1);
+        if (next != '\\')
+          break;
+      }
+      result += next;
+    }
+
+    result.shrink_to_fit();
+    return result;
+  }
+
+  // Returns a std::vector containing all remaining bytes of the input data.
+  template <typename T> std::vector<T> ConsumeRemainingBytes() {
+    return ConsumeBytes<T>(remaining_bytes_);
+  }
+
+  // Returns a std::string containing all remaining bytes of the input data.
+  // Prefer using |ConsumeRemainingBytes| unless you actually need a std::string
+  // object.
+  std::string ConsumeRemainingBytesAsString() {
+    return ConsumeBytesAsString(remaining_bytes_);
+  }
+
+  // Returns a number in the range [Type's min, Type's max]. The value might
+  // not be uniformly distributed in the given range. If there's no input data
+  // left, always returns |min|.
+  template <typename T> T ConsumeIntegral() {
+    return ConsumeIntegralInRange(std::numeric_limits<T>::min(),
+                                  std::numeric_limits<T>::max());
+  }
+
+  // Reads one byte and returns a bool, or false when no data remains.
+  bool ConsumeBool() { return 1 & ConsumeIntegral<uint8_t>(); }
+
+  // Returns a copy of the value selected from the given fixed-size |array|.
+  template <typename T, size_t size>
+  T PickValueInArray(const T (&array)[size]) {
+    static_assert(size > 0, "The array must be non empty.");
+    return array[ConsumeIntegralInRange<size_t>(0, size - 1)];
+  }
+
+  template <typename T>
+  T PickValueInArray(std::initializer_list<const T> list) {
+    // TODO(Dor1s): switch to static_assert once C++14 is allowed.
+    if (!list.size())
+      abort();
+
+    return *(list.begin() + ConsumeIntegralInRange<size_t>(0, list.size() - 1));
+  }
+
+  // Returns an enum value. The enum must start at 0 and be contiguous. It must
+  // also contain |kMaxValue| aliased to its largest (inclusive) value. Such as:
+  // enum class Foo { SomeValue, OtherValue, kMaxValue = OtherValue };
+  template <typename T> T ConsumeEnum() {
+    static_assert(std::is_enum<T>::value, "|T| must be an enum type.");
+    return static_cast<T>(ConsumeIntegralInRange<uint32_t>(
+        0, static_cast<uint32_t>(T::kMaxValue)));
+  }
+
+  // Returns a floating point number in the range [0.0, 1.0]. If there's no
+  // input data left, always returns 0.
+  template <typename T> T ConsumeProbability() {
+    static_assert(std::is_floating_point<T>::value,
+                  "A floating point type is required.");
+
+    // Use different integral types for different floating point types in order
+    // to provide better density of the resulting values.
+    using IntegralType =
+        typename std::conditional<sizeof(T) <= sizeof(uint32_t), uint32_t,
+                                  uint64_t>::type;
+
+    T result = static_cast<T>(ConsumeIntegral<IntegralType>());
+    result /= static_cast<T>(std::numeric_limits<IntegralType>::max());
+    return result;
+  }
+
+  // Returns a floating point value in the range [Type's lowest, Type's max] by
+  // consuming bytes from the input data. If there's no input data left, always
+  // returns approximately 0.
+  template <typename T> T ConsumeFloatingPoint() {
+    return ConsumeFloatingPointInRange<T>(std::numeric_limits<T>::lowest(),
+                                          std::numeric_limits<T>::max());
+  }
+
+  // Returns a floating point value in the given range by consuming bytes from
+  // the input data. If there's no input data left, returns |min|. Note that
+  // |min| must be less than or equal to |max|.
+  template <typename T> T ConsumeFloatingPointInRange(T min, T max) {
+    if (min > max)
+      abort();
+
+    T range = .0;
+    T result = min;
+    constexpr T zero(.0);
+    if (max > zero && min < zero && max > min + std::numeric_limits<T>::max()) {
+      // The diff |max - min| would overflow the given floating point type. Use
+      // the half of the diff as the range and consume a bool to decide whether
+      // the result is in the first of the second part of the diff.
+      range = (max / 2.0) - (min / 2.0);
+      if (ConsumeBool()) {
+        result += range;
+      }
+    } else {
+      range = max - min;
+    }
+
+    return result + range * ConsumeProbability<T>();
+  }
+
+  // Reports the remaining bytes available for fuzzed input.
+  size_t remaining_bytes() { return remaining_bytes_; }
+
+ private:
+  FuzzedDataProvider(const FuzzedDataProvider &) = delete;
+  FuzzedDataProvider &operator=(const FuzzedDataProvider &) = delete;
+
+  void Advance(size_t num_bytes) {
+    if (num_bytes > remaining_bytes_)
+      abort();
+
+    data_ptr_ += num_bytes;
+    remaining_bytes_ -= num_bytes;
+  }
+
+  template <typename T>
+  std::vector<T> ConsumeBytes(size_t size, size_t num_bytes_to_consume) {
+    static_assert(sizeof(T) == sizeof(uint8_t), "Incompatible data type.");
+
+    // The point of using the size-based constructor below is to increase the
+    // odds of having a vector object with capacity being equal to the length.
+    // That part is always implementation specific, but at least both libc++ and
+    // libstdc++ allocate the requested number of bytes in that constructor,
+    // which seems to be a natural choice for other implementations as well.
+    // To increase the odds even more, we also call |shrink_to_fit| below.
+    std::vector<T> result(size);
+    std::memcpy(result.data(), data_ptr_, num_bytes_to_consume);
+    Advance(num_bytes_to_consume);
+
+    // Even though |shrink_to_fit| is also implementation specific, we expect it
+    // to provide an additional assurance in case vector's constructor allocated
+    // a buffer which is larger than the actual amount of data we put inside it.
+    result.shrink_to_fit();
+    return result;
+  }
+
+  template <typename TS, typename TU> TS ConvertUnsignedToSigned(TU value) {
+    static_assert(sizeof(TS) == sizeof(TU), "Incompatible data types.");
+    static_assert(!std::numeric_limits<TU>::is_signed,
+                  "Source type must be unsigned.");
+
+    // TODO(Dor1s): change to `if constexpr` once C++17 becomes mainstream.
+    if (std::numeric_limits<TS>::is_modulo)
+      return static_cast<TS>(value);
+
+    // Avoid using implementation-defined unsigned to signer conversions.
+    // To learn more, see https://stackoverflow.com/questions/13150449.
+    if (value <= std::numeric_limits<TS>::max())
+      return static_cast<TS>(value);
+    else {
+      constexpr auto TS_min = std::numeric_limits<TS>::min();
+      return TS_min + static_cast<char>(value - TS_min);
+    }
+  }
+
+  const uint8_t *data_ptr_;
+  size_t remaining_bytes_;
+};
+
+#endif // LLVM_FUZZER_FUZZED_DATA_PROVIDER_H_
diff --git a/linux-x86/lib64/clang/9.0.7/include/fxsrintrin.h b/linux-x86/lib64/clang/10.0.1/include/fxsrintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/fxsrintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/fxsrintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/gfniintrin.h b/linux-x86/lib64/clang/10.0.1/include/gfniintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/gfniintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/gfniintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/htmintrin.h b/linux-x86/lib64/clang/10.0.1/include/htmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/htmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/htmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/htmxlintrin.h b/linux-x86/lib64/clang/10.0.1/include/htmxlintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/htmxlintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/htmxlintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/ia32intrin.h b/linux-x86/lib64/clang/10.0.1/include/ia32intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/ia32intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/ia32intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/immintrin.h b/linux-x86/lib64/clang/10.0.1/include/immintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/immintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/immintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/intrin.h b/linux-x86/lib64/clang/10.0.1/include/intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/inttypes.h b/linux-x86/lib64/clang/10.0.1/include/inttypes.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/inttypes.h
rename to linux-x86/lib64/clang/10.0.1/include/inttypes.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/invpcidintrin.h b/linux-x86/lib64/clang/10.0.1/include/invpcidintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/invpcidintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/invpcidintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/iso646.h b/linux-x86/lib64/clang/10.0.1/include/iso646.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/iso646.h
rename to linux-x86/lib64/clang/10.0.1/include/iso646.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/limits.h b/linux-x86/lib64/clang/10.0.1/include/limits.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/limits.h
rename to linux-x86/lib64/clang/10.0.1/include/limits.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/lwpintrin.h b/linux-x86/lib64/clang/10.0.1/include/lwpintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/lwpintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/lwpintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/lzcntintrin.h b/linux-x86/lib64/clang/10.0.1/include/lzcntintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/lzcntintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/lzcntintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/mm3dnow.h b/linux-x86/lib64/clang/10.0.1/include/mm3dnow.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/mm3dnow.h
rename to linux-x86/lib64/clang/10.0.1/include/mm3dnow.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/mm_malloc.h b/linux-x86/lib64/clang/10.0.1/include/mm_malloc.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/mm_malloc.h
rename to linux-x86/lib64/clang/10.0.1/include/mm_malloc.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/mmintrin.h b/linux-x86/lib64/clang/10.0.1/include/mmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/mmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/mmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/module.modulemap b/linux-x86/lib64/clang/10.0.1/include/module.modulemap
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/module.modulemap
rename to linux-x86/lib64/clang/10.0.1/include/module.modulemap
diff --git a/linux-x86/lib64/clang/9.0.7/include/movdirintrin.h b/linux-x86/lib64/clang/10.0.1/include/movdirintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/movdirintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/movdirintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/msa.h b/linux-x86/lib64/clang/10.0.1/include/msa.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/msa.h
rename to linux-x86/lib64/clang/10.0.1/include/msa.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/mwaitxintrin.h b/linux-x86/lib64/clang/10.0.1/include/mwaitxintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/mwaitxintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/mwaitxintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/nmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/nmmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/nmmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/nmmintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/omp.h b/linux-x86/lib64/clang/10.0.1/include/omp.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/omp.h
copy to linux-x86/lib64/clang/10.0.1/include/omp.h
index 7ba1859..9c37698 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/omp.h
+++ b/linux-x86/lib64/clang/10.0.1/include/omp.h
@@ -1,5 +1,5 @@
 /*
- * include/50/omp.h.var
+ * include/omp.h.var
  */
 
 
@@ -15,6 +15,9 @@
 #ifndef __OMP_H
 #   define __OMP_H
 
+#   include <stdlib.h>
+#   include <stdint.h>
+
 #   define KMP_VERSION_MAJOR    5
 #   define KMP_VERSION_MINOR    0
 #   define KMP_VERSION_BUILD    20140926
@@ -134,8 +137,6 @@
     extern int  __KAI_KMPC_CONVENTION  omp_get_team_num (void);
     extern int  __KAI_KMPC_CONVENTION  omp_get_cancellation (void);
 
-#   include <stdlib.h>
-#   include <stdint.h>
     /* OpenMP 4.5 */
     extern int   __KAI_KMPC_CONVENTION  omp_get_initial_device (void);
     extern void* __KAI_KMPC_CONVENTION  omp_target_alloc(size_t, int);
diff --git a/darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h b/linux-x86/lib64/clang/10.0.1/include/opencl-c-base.h
similarity index 97%
copy from darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h
copy to linux-x86/lib64/clang/10.0.1/include/opencl-c-base.h
index a82954d..430e07d 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/opencl-c-base.h
+++ b/linux-x86/lib64/clang/10.0.1/include/opencl-c-base.h
@@ -126,7 +126,7 @@
 typedef double double16 __attribute__((ext_vector_type(16)));
 #endif
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #define NULL ((void*)0)
 #endif
 
@@ -276,7 +276,7 @@
  */
 #define CLK_GLOBAL_MEM_FENCE   0x02
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 typedef enum memory_scope {
   memory_scope_work_item = __OPENCL_MEMORY_SCOPE_WORK_ITEM,
@@ -288,9 +288,6 @@
 #endif
 } memory_scope;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
 /**
  * Queue a memory fence to ensure correct ordering of memory
  * operations between work-items of a work-group to
@@ -313,7 +310,7 @@
   memory_order_seq_cst = __ATOMIC_SEQ_CST
 } memory_order;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.3, v1.2 s6.12.14, v2.0 s6.13.14 - Image Read and Write Functions
 
@@ -389,14 +386,10 @@
 #endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
 
 // OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #define CLK_NULL_RESERVE_ID (__builtin_astype(((void*)(__SIZE_MAX__)), reserve_id_t))
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
 
 // OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
 #define CL_COMPLETE                                 0x0
 #define CL_RUNNING                                  0x1
 #define CL_SUBMITTED                                0x2
@@ -413,7 +406,7 @@
 #define CLK_OUT_OF_RESOURCES                        -5
 
 #define CLK_NULL_QUEUE                              0
-#define CLK_NULL_EVENT (__builtin_astype(((void*)(__SIZE_MAX__)), clk_event_t))
+#define CLK_NULL_EVENT (__builtin_astype(((__SIZE_MAX__)), clk_event_t))
 
 // execution model related definitions
 #define CLK_ENQUEUE_FLAGS_NO_WAIT                   0x0
@@ -435,7 +428,7 @@
   size_t localWorkSize[MAX_WORK_DIM];
 } ndrange_t;
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 #ifdef cl_intel_device_side_avc_motion_estimation
 #pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : begin
diff --git a/darwin-x86/lib64/clang/9.0.7/include/opencl-c.h b/linux-x86/lib64/clang/10.0.1/include/opencl-c.h
similarity index 98%
copy from darwin-x86/lib64/clang/9.0.7/include/opencl-c.h
copy to linux-x86/lib64/clang/10.0.1/include/opencl-c.h
index 4207c53..06c5ab6 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/opencl-c.h
+++ b/linux-x86/lib64/clang/10.0.1/include/opencl-c.h
@@ -11,11 +11,11 @@
 
 #include "opencl-c-base.h"
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifndef cl_khr_depth_images
 #define cl_khr_depth_images
 #endif //cl_khr_depth_images
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 #if __OPENCL_C_VERSION__ < CL_VERSION_2_0
 #ifdef cl_khr_3d_image_writes
@@ -23,10 +23,10 @@
 #endif //cl_khr_3d_image_writes
 #endif //__OPENCL_C_VERSION__ < CL_VERSION_2_0
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 #pragma OPENCL EXTENSION cl_intel_planar_yuv : begin
 #pragma OPENCL EXTENSION cl_intel_planar_yuv : end
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 #define __ovld __attribute__((overloadable))
 #define __conv __attribute__((convergent))
@@ -6517,11 +6517,11 @@
  */
 size_t __ovld __cnfn get_global_offset(uint dimindx);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 size_t __ovld get_enqueued_local_size(uint dimindx);
 size_t __ovld get_global_linear_id(void);
 size_t __ovld get_local_linear_id(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.2, v1.2 s6.12.2, v2.0 s6.13.2 - Math functions
 
@@ -7352,7 +7352,7 @@
  * Returns fmin(x - floor (x), 0x1.fffffep-1f ).
  * floor(x) is returned in iptr.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld fract(float x, float *iptr);
 float2 __ovld fract(float2 x, float2 *iptr);
 float3 __ovld fract(float3 x, float3 *iptr);
@@ -7434,7 +7434,7 @@
 half8 __ovld fract(half8 x, __private half8 *iptr);
 half16 __ovld fract(half16 x, __private half16 *iptr);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Extract mantissa and exponent from x. For each
@@ -7442,7 +7442,7 @@
  * magnitude in the interval [1/2, 1) or 0. Each
  * component of x equals mantissa returned * 2^exp.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld frexp(float x, int *exp);
 float2 __ovld frexp(float2 x, int2 *exp);
 float3 __ovld frexp(float3 x, int3 *exp);
@@ -7524,7 +7524,7 @@
 half8 __ovld frexp(half8 x, __private int8 *exp);
 half16 __ovld frexp(half16 x, __private int16 *exp);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute the value of the square root of x^2 + y^2
@@ -7649,7 +7649,7 @@
 half16 __ovld __cnfn lgamma(half16 x);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld lgamma_r(float x, int *signp);
 float2 __ovld lgamma_r(float2 x, int2 *signp);
 float3 __ovld lgamma_r(float3 x, int3 *signp);
@@ -7731,7 +7731,7 @@
 half8 __ovld lgamma_r(half8 x, __private int8 *signp);
 half16 __ovld lgamma_r(half16 x, __private int16 *signp);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute natural logarithm.
@@ -7955,7 +7955,7 @@
  * the argument. It stores the integral part in the object
  * pointed to by iptr.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld modf(float x, float *iptr);
 float2 __ovld modf(float2 x, float2 *iptr);
 float3 __ovld modf(float3 x, float3 *iptr);
@@ -8037,7 +8037,7 @@
 half8 __ovld modf(half8 x, __private half8 *iptr);
 half16 __ovld modf(half16 x, __private half16 *iptr);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Returns a quiet NaN. The nancode may be placed
@@ -8215,7 +8215,7 @@
  * sign as x/y. It stores this signed value in the object
  * pointed to by quo.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld remquo(float x, float y, int *quo);
 float2 __ovld remquo(float2 x, float2 y, int2 *quo);
 float3 __ovld remquo(float3 x, float3 y, int3 *quo);
@@ -8298,7 +8298,7 @@
 half8 __ovld remquo(half8 x, half8 y, __private int8 *quo);
 half16 __ovld remquo(half16 x, half16 y, __private int16 *quo);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 /**
  * Round to integral value (using round to nearest
  * even rounding mode) in floating-point format.
@@ -8439,7 +8439,7 @@
  * is the return value and computed cosine is returned
  * in cosval.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld sincos(float x, float *cosval);
 float2 __ovld sincos(float2 x, float2 *cosval);
 float3 __ovld sincos(float3 x, float3 *cosval);
@@ -8521,7 +8521,7 @@
 half8 __ovld sincos(half8 x, __private half8 *cosval);
 half16 __ovld sincos(half16 x, __private half16 *cosval);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Compute hyperbolic sine.
@@ -9446,7 +9446,7 @@
  * returns the size in bits of the type of x or
  * component type of x, if x is a vector.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 char __ovld ctz(char x);
 uchar __ovld ctz(uchar x);
 char2 __ovld ctz(char2 x);
@@ -9495,7 +9495,7 @@
 ulong8 __ovld ctz(ulong8 x);
 long16 __ovld ctz(long16 x);
 ulong16 __ovld ctz(ulong16 x);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Returns mul_hi(a, b) + c.
@@ -11340,7 +11340,7 @@
 half16 __ovld vload16(size_t offset, const __constant half *p);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 char2 __ovld vload2(size_t offset, const char *p);
 uchar2 __ovld vload2(size_t offset, const uchar *p);
 short2 __ovld vload2(size_t offset, const short *p);
@@ -11578,9 +11578,9 @@
 half8 __ovld vload8(size_t offset, const __private half *p);
 half16 __ovld vload16(size_t offset, const __private half *p);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore2(char2 data, size_t offset, char *p);
 void __ovld vstore2(uchar2 data, size_t offset, uchar *p);
 void __ovld vstore2(short2 data, size_t offset, short *p);
@@ -11814,7 +11814,7 @@
 void __ovld vstore8(half8 data, size_t offset, __private half *p);
 void __ovld vstore16(half16 data, size_t offset, __private half *p);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Read sizeof (half) bytes of data from address
@@ -11825,13 +11825,13 @@
  * must be 16-bit aligned.
  */
 float __ovld vload_half(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld vload_half(size_t offset, const half *p);
 #else
 float __ovld vload_half(size_t offset, const __global half *p);
 float __ovld vload_half(size_t offset, const __local half *p);
 float __ovld vload_half(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Read sizeof (halfn) bytes of data from address
@@ -11846,7 +11846,7 @@
 float4 __ovld vload_half4(size_t offset, const __constant half *p);
 float8 __ovld vload_half8(size_t offset, const __constant half *p);
 float16 __ovld vload_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float2 __ovld vload_half2(size_t offset, const half *p);
 float3 __ovld vload_half3(size_t offset, const half *p);
 float4 __ovld vload_half4(size_t offset, const half *p);
@@ -11868,7 +11868,7 @@
 float4 __ovld vload_half4(size_t offset, const __private half *p);
 float8 __ovld vload_half8(size_t offset, const __private half *p);
 float16 __ovld vload_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The float value given by data is first
@@ -11881,7 +11881,7 @@
  * The default current rounding mode is round to
  * nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore_half(float data, size_t offset, half *p);
 void __ovld vstore_half_rte(float data, size_t offset, half *p);
 void __ovld vstore_half_rtz(float data, size_t offset, half *p);
@@ -11927,7 +11927,7 @@
 void __ovld vstore_half_rtp(double data, size_t offset, __private half *p);
 void __ovld vstore_half_rtn(double data, size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The floatn value given by data is converted to
@@ -11940,7 +11940,7 @@
  * The default current rounding mode is round to
  * nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstore_half2(float2 data, size_t offset, half *p);
 void __ovld vstore_half3(float3 data, size_t offset, half *p);
 void __ovld vstore_half4(float4 data, size_t offset, half *p);
@@ -12146,7 +12146,7 @@
 void __ovld vstore_half8_rtn(double8 data, size_t offset, __private half *p);
 void __ovld vstore_half16_rtn(double16 data, size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * For n = 1, 2, 4, 8 and 16 read sizeof (halfn)
@@ -12167,7 +12167,7 @@
 float4 __ovld vloada_half4(size_t offset, const __constant half *p);
 float8 __ovld vloada_half8(size_t offset, const __constant half *p);
 float16 __ovld vloada_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float __ovld vloada_half(size_t offset, const half *p);
 float2 __ovld vloada_half2(size_t offset, const half *p);
 float3 __ovld vloada_half3(size_t offset, const half *p);
@@ -12193,7 +12193,7 @@
 float4 __ovld vloada_half4(size_t offset, const __private half *p);
 float8 __ovld vloada_half8(size_t offset, const __private half *p);
 float16 __ovld vloada_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * The floatn value given by data is converted to
@@ -12211,7 +12211,7 @@
  * mode. The default current rounding mode is
  * round to nearest even.
  */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld vstorea_half(float data, size_t offset, half *p);
 void __ovld vstorea_half2(float2 data, size_t offset, half *p);
 void __ovld vstorea_half3(float3 data, size_t offset, half *p);
@@ -12496,7 +12496,7 @@
 void __ovld vstorea_half8_rtn(double8 data,size_t offset, __private half *p);
 void __ovld vstorea_half16_rtn(double16 data,size_t offset, __private half *p);
 #endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.8, v1.2 s6.12.8, v2.0 s6.13.8 - Synchronization Functions
 
@@ -12532,10 +12532,10 @@
 
 void __ovld __conv barrier(cl_mem_fence_flags flags);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld __conv work_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
 void __ovld __conv work_group_barrier(cl_mem_fence_flags flags);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.9, v1.2 s6.12.9 - Explicit Memory Fence Functions
 
@@ -12580,7 +12580,7 @@
 
 // OpenCL v2.0 s6.13.9 - Address Space Qualifier Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 cl_mem_fence_flags __ovld get_fence(const void *ptr);
 cl_mem_fence_flags __ovld get_fence(void *ptr);
 
@@ -12591,7 +12591,7 @@
  * where gentype is builtin type or user defined type.
  */
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.10, v1.2 s6.12.10, v2.0 s6.13.10 - Async Copies from Global to Local Memory, Local to Global Memory, and Prefetch
 
@@ -13371,7 +13371,7 @@
 
 // OpenCL v2.0 s6.13.11 - Atomics Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // double atomics support requires extensions cl_khr_int64_base_atomics and cl_khr_int64_extended_atomics
 #if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
@@ -13692,7 +13692,7 @@
 void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order);
 void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order, memory_scope scope);
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v1.1 s6.11.12, v1.2 s6.12.12, v2.0 s6.13.12 - Miscellaneous Vector Functions
 
@@ -14186,7 +14186,7 @@
 half16 __ovld __cnfn shuffle2(half16 x, half16 y, ushort16 mask);
 #endif //cl_khr_fp16
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 // OpenCL v1.2 s6.12.13, v2.0 s6.13.13 - printf
 
 int printf(__constant const char* st, ...) __attribute__((format(printf, 1, 2)));
@@ -14307,7 +14307,7 @@
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
 float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
 
@@ -14315,7 +14315,7 @@
 int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
 uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, int coord);
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord);
@@ -14325,7 +14325,7 @@
 uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, int coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, float coord);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
 float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
 
@@ -14333,7 +14333,7 @@
 int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
 uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 #ifdef cl_khr_depth_images
 float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, float2 coord);
@@ -14358,7 +14358,7 @@
 #endif //cl_khr_gl_msaa_sharing
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 
 float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord, float lod);
@@ -14410,9 +14410,9 @@
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 /**
 * Sampler-less Image Access
@@ -14447,7 +14447,7 @@
 int4 __purefn __ovld read_imagei(read_only image3d_t image, int4 coord);
 uint4 __purefn __ovld read_imageui(read_only image3d_t image, int4 coord);
 
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 
 // Image read functions returning half4 type
 #ifdef cl_khr_fp16
@@ -14457,7 +14457,7 @@
 half4 __purefn __ovld read_imageh(read_only image2d_t image, sampler_t sampler, float2 coord);
 half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, int4 coord);
 half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, float4 coord);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, int2 coord);
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, float2 coord);
 half4 __purefn __ovld read_imageh(read_only image2d_array_t image, sampler_t sampler, int4 coord);
@@ -14471,11 +14471,11 @@
 half4 __purefn __ovld read_imageh(read_only image1d_array_t image, int2 coord);
 half4 __purefn __ovld read_imageh(read_only image2d_array_t image, int4 coord);
 half4 __purefn __ovld read_imageh(read_only image1d_buffer_t image, int coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_1_2)
 #endif //cl_khr_fp16
 
 // Image read functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 float4 __purefn __ovld read_imagef(read_write image1d_t image, int coord);
 int4 __purefn __ovld read_imagei(read_write image1d_t image, int coord);
 uint4 __purefn __ovld read_imageui(read_write image1d_t image, int coord);
@@ -14518,7 +14518,7 @@
 float __purefn __ovld read_imagef(read_write image2d_array_msaa_depth_t image, int4 coord, int sample);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 float4 __purefn __ovld read_imagef(read_write image1d_t image, sampler_t sampler, float coord, float lod);
 int4 __purefn __ovld read_imagei(read_write image1d_t image, sampler_t sampler, float coord, float lod);
@@ -14569,7 +14569,7 @@
 uint4 __purefn __ovld read_imageui(read_write image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image read functions returning half4 type
 #ifdef cl_khr_fp16
@@ -14580,7 +14580,7 @@
 half4 __purefn __ovld read_imageh(read_write image2d_array_t image, int4 coord);
 half4 __purefn __ovld read_imageh(read_write image1d_buffer_t image, int coord);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Write color value to location specified by coordinate
@@ -14681,7 +14681,7 @@
 #endif //cl_khr_depth_images
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 void __ovld write_imagef(write_only image1d_t image, int coord, int lod, float4 color);
 void __ovld write_imagei(write_only image1d_t image, int coord, int lod, int4 color);
@@ -14708,7 +14708,7 @@
 void __ovld write_imageui(write_only image3d_t image, int4 coord, int lod, uint4 color);
 #endif
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image write functions for half4 type
 #ifdef cl_khr_fp16
@@ -14723,7 +14723,7 @@
 #endif //cl_khr_fp16
 
 // Image write functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void __ovld write_imagef(read_write image2d_t image, int2 coord, float4 color);
 void __ovld write_imagei(read_write image2d_t image, int2 coord, int4 color);
 void __ovld write_imageui(read_write image2d_t image, int2 coord, uint4 color);
@@ -14755,7 +14755,7 @@
 void __ovld write_imagef(read_write image2d_array_depth_t image, int4 coord, float color);
 #endif //cl_khr_depth_images
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 void __ovld write_imagef(read_write image1d_t image, int coord, int lod, float4 color);
 void __ovld write_imagei(read_write image1d_t image, int coord, int lod, int4 color);
@@ -14782,7 +14782,7 @@
 void __ovld write_imageui(read_write image3d_t image, int4 coord, int lod, uint4 color);
 #endif
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Image write functions for half4 type
 #ifdef cl_khr_fp16
@@ -14795,7 +14795,7 @@
 void __ovld write_imageh(read_write image2d_array_t image, int4 coord, half4 color);
 void __ovld write_imageh(read_write image1d_buffer_t image, int coord, half4 color);
 #endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // Note: In OpenCL v1.0/1.1/1.2, image argument of image query builtin functions does not have
 // access qualifier, which by default assume read_only access qualifier. Image query builtin
@@ -14843,7 +14843,7 @@
 int __ovld __cnfn get_image_width(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_width(read_write image1d_t image);
 int __ovld __cnfn get_image_width(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_width(read_write image2d_t image);
@@ -14860,7 +14860,7 @@
 int __ovld __cnfn get_image_width(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_width(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image height in pixels.
@@ -14895,7 +14895,7 @@
 int __ovld __cnfn get_image_height(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_height(read_write image2d_t image);
 int __ovld __cnfn get_image_height(read_write image3d_t image);
 int __ovld __cnfn get_image_height(read_write image2d_array_t image);
@@ -14909,7 +14909,7 @@
 int __ovld __cnfn get_image_height(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_height(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image depth in pixels.
@@ -14920,12 +14920,12 @@
 int __ovld __cnfn get_image_depth(write_only image3d_t image);
 #endif
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_depth(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #ifdef cl_khr_mipmap_image
 /**
  * Return the image miplevels.
@@ -14961,7 +14961,7 @@
 int __ovld get_image_num_mip_levels(read_write image2d_depth_t image);
 
 #endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the channel data type. Valid values are:
@@ -15018,7 +15018,7 @@
 int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_channel_data_type(read_write image1d_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_t image);
@@ -15035,7 +15035,7 @@
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image channel order. Valid values are:
@@ -15090,7 +15090,7 @@
 int __ovld __cnfn get_image_channel_order(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __cnfn get_image_channel_order(read_write image1d_t image);
 int __ovld __cnfn get_image_channel_order(read_write image1d_buffer_t image);
 int __ovld __cnfn get_image_channel_order(read_write image2d_t image);
@@ -15107,7 +15107,7 @@
 int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_t image);
 int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the 2D image width and height as an int2
@@ -15140,7 +15140,7 @@
 int2 __ovld __cnfn get_image_dim(write_only image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int2 __ovld __cnfn get_image_dim(read_write image2d_t image);
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_t image);
 #ifdef cl_khr_depth_images
@@ -15153,7 +15153,7 @@
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_t image);
 int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_depth_t image);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the 3D image width, height, and depth as an
@@ -15165,9 +15165,9 @@
 #ifdef cl_khr_3d_image_writes
 int4 __ovld __cnfn get_image_dim(write_only image3d_t image);
 #endif
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int4 __ovld __cnfn get_image_dim(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
  * Return the image array size.
@@ -15193,7 +15193,7 @@
 size_t __ovld __cnfn get_image_array_size(write_only image2d_array_msaa_depth_t image_array);
 #endif //cl_khr_gl_msaa_sharing
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 size_t __ovld __cnfn get_image_array_size(read_write image1d_array_t image_array);
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_t image_array);
 #ifdef cl_khr_depth_images
@@ -15203,7 +15203,7 @@
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_t image_array);
 size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_depth_t image_array);
 #endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 /**
 * Return the number of samples associated with image
@@ -15219,17 +15219,17 @@
 int __ovld get_image_num_samples(write_only image2d_array_msaa_t image);
 int __ovld get_image_num_samples(write_only image2d_array_msaa_depth_t image);
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld get_image_num_samples(read_write image2d_msaa_t image);
 int __ovld get_image_num_samples(read_write image2d_msaa_depth_t image);
 int __ovld get_image_num_samples(read_write image2d_array_msaa_t image);
 int __ovld get_image_num_samples(read_write image2d_array_msaa_depth_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 #endif
 
 // OpenCL v2.0 s6.13.15 - Work-group Functions
 
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 int __ovld __conv work_group_all(int predicate);
 int __ovld __conv work_group_any(int predicate);
 
@@ -15327,16 +15327,16 @@
 double __ovld __conv work_group_scan_inclusive_max(double x);
 #endif //cl_khr_fp64
 
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 bool __ovld is_valid_reserve_id(reserve_id_t reserve_id);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 
 // OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 ndrange_t __ovld ndrange_1D(size_t);
 ndrange_t __ovld ndrange_1D(size_t, size_t);
@@ -15350,7 +15350,7 @@
 ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3]);
 ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3], const size_t[3]);
 
-int __ovld enqueue_marker(queue_t, uint, const __private clk_event_t*, __private clk_event_t*);
+int __ovld enqueue_marker(queue_t, uint, const clk_event_t*, clk_event_t*);
 
 void __ovld retain_event(clk_event_t);
 
@@ -15365,7 +15365,7 @@
 void __ovld capture_event_profiling_info(clk_event_t, clk_profiling_info, __global void* value);
 
 queue_t __ovld get_default_queue(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 // OpenCL Extension v2.0 s9.17 - Sub-groups
 
@@ -15374,16 +15374,16 @@
 uint    __ovld get_sub_group_size(void);
 uint    __ovld get_max_sub_group_size(void);
 uint    __ovld get_num_sub_groups(void);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld get_enqueued_num_sub_groups(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld get_sub_group_id(void);
 uint    __ovld get_sub_group_local_id(void);
 
 void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
+#endif //defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 int     __ovld __conv sub_group_all(int predicate);
 int     __ovld __conv sub_group_any(int predicate);
@@ -15573,12 +15573,12 @@
 uint4   __ovld __conv intel_sub_group_block_read4( read_only image2d_t image, int2 coord );
 uint8   __ovld __conv intel_sub_group_block_read8( read_only image2d_t image, int2 coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint    __ovld __conv intel_sub_group_block_read(read_write image2d_t image, int2 coord);
 uint2   __ovld __conv intel_sub_group_block_read2(read_write image2d_t image, int2 coord);
 uint4   __ovld __conv intel_sub_group_block_read4(read_write image2d_t image, int2 coord);
 uint8   __ovld __conv intel_sub_group_block_read8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 uint    __ovld __conv intel_sub_group_block_read( const __global uint* p );
 uint2   __ovld __conv intel_sub_group_block_read2( const __global uint* p );
@@ -15590,12 +15590,12 @@
 void    __ovld __conv intel_sub_group_block_write4(write_only image2d_t image, int2 coord, uint4 data);
 void    __ovld __conv intel_sub_group_block_write8(write_only image2d_t image, int2 coord, uint8 data);
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void    __ovld __conv intel_sub_group_block_write(read_write image2d_t image, int2 coord, uint data);
 void    __ovld __conv intel_sub_group_block_write2(read_write image2d_t image, int2 coord, uint2 data);
 void    __ovld __conv intel_sub_group_block_write4(read_write image2d_t image, int2 coord, uint4 data);
 void    __ovld __conv intel_sub_group_block_write8(read_write image2d_t image, int2 coord, uint8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void    __ovld __conv intel_sub_group_block_write( __global uint* p, uint data );
 void    __ovld __conv intel_sub_group_block_write2( __global uint* p, uint2 data );
@@ -15713,12 +15713,12 @@
 uint4      __ovld __conv intel_sub_group_block_read_ui4( read_only image2d_t image, int2 byte_coord );
 uint8      __ovld __conv intel_sub_group_block_read_ui8( read_only image2d_t image, int2 byte_coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 uint       __ovld __conv intel_sub_group_block_read_ui( read_write image2d_t image, int2 byte_coord );
 uint2      __ovld __conv intel_sub_group_block_read_ui2( read_write image2d_t image, int2 byte_coord );
 uint4      __ovld __conv intel_sub_group_block_read_ui4( read_write image2d_t image, int2 byte_coord );
 uint8      __ovld __conv intel_sub_group_block_read_ui8( read_write image2d_t image, int2 byte_coord );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 uint       __ovld __conv intel_sub_group_block_read_ui( const __global uint* p );
 uint2      __ovld __conv intel_sub_group_block_read_ui2( const __global uint* p );
@@ -15730,12 +15730,12 @@
 void       __ovld __conv intel_sub_group_block_write_ui4( read_only image2d_t image, int2 byte_coord, uint4 data );
 void       __ovld __conv intel_sub_group_block_write_ui8( read_only image2d_t image, int2 byte_coord, uint8 data );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void       __ovld __conv intel_sub_group_block_write_ui( read_write image2d_t image, int2 byte_coord, uint data );
 void       __ovld __conv intel_sub_group_block_write_ui2( read_write image2d_t image, int2 byte_coord, uint2 data );
 void       __ovld __conv intel_sub_group_block_write_ui4( read_write image2d_t image, int2 byte_coord, uint4 data );
 void       __ovld __conv intel_sub_group_block_write_ui8( read_write image2d_t image, int2 byte_coord, uint8 data );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void       __ovld __conv intel_sub_group_block_write_ui( __global uint* p, uint data );
 void       __ovld __conv intel_sub_group_block_write_ui2( __global uint* p, uint2 data );
@@ -15747,12 +15747,12 @@
 ushort4     __ovld __conv intel_sub_group_block_read_us4( read_only image2d_t image, int2 coord );
 ushort8     __ovld __conv intel_sub_group_block_read_us8( read_only image2d_t image, int2 coord );
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 ushort      __ovld __conv intel_sub_group_block_read_us(read_write image2d_t image, int2 coord);
 ushort2     __ovld __conv intel_sub_group_block_read_us2(read_write image2d_t image, int2 coord);
 ushort4     __ovld __conv intel_sub_group_block_read_us4(read_write image2d_t image, int2 coord);
 ushort8     __ovld __conv intel_sub_group_block_read_us8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 ushort      __ovld __conv intel_sub_group_block_read_us(  const __global ushort* p );
 ushort2     __ovld __conv intel_sub_group_block_read_us2( const __global ushort* p );
@@ -15764,12 +15764,12 @@
 void        __ovld __conv intel_sub_group_block_write_us4(write_only image2d_t image, int2 coord, ushort4 data);
 void        __ovld __conv intel_sub_group_block_write_us8(write_only image2d_t image, int2 coord, ushort8 data);
 
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#if defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 void        __ovld __conv intel_sub_group_block_write_us(read_write image2d_t image, int2 coord, ushort  data);
 void        __ovld __conv intel_sub_group_block_write_us2(read_write image2d_t image, int2 coord, ushort2 data);
 void        __ovld __conv intel_sub_group_block_write_us4(read_write image2d_t image, int2 coord, ushort4 data);
 void        __ovld __conv intel_sub_group_block_write_us8(read_write image2d_t image, int2 coord, ushort8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
+#endif // defined(__OPENCL_CPP_VERSION__) || (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
 
 void        __ovld __conv intel_sub_group_block_write_us(  __global ushort* p, ushort  data );
 void        __ovld __conv intel_sub_group_block_write_us2( __global ushort* p, ushort2 data );
diff --git a/linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math.h b/linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math.h
rename to linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math_declares.h b/linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math_declares.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/__clang_openmp_math_declares.h
rename to linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/__clang_openmp_math_declares.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/cmath b/linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/cmath
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/cmath
rename to linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/cmath
diff --git a/linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/math.h b/linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/math.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/openmp_wrappers/math.h
rename to linux-x86/lib64/clang/10.0.1/include/openmp_wrappers/math.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/pconfigintrin.h b/linux-x86/lib64/clang/10.0.1/include/pconfigintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/pconfigintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/pconfigintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/pkuintrin.h b/linux-x86/lib64/clang/10.0.1/include/pkuintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/pkuintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/pkuintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/pmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/pmmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/pmmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/pmmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/popcntintrin.h b/linux-x86/lib64/clang/10.0.1/include/popcntintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/popcntintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/popcntintrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
index 617ce24..293276c 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/emmintrin.h
@@ -35,6 +35,8 @@
 #ifndef EMMINTRIN_H_
 #define EMMINTRIN_H_
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <altivec.h>
 
 /* We need definitions from the SSE header files.  */
@@ -2315,4 +2317,8 @@
   return (__m128d) __A;
 }
 
+#else
+#include_next <emmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* EMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
similarity index 84%
copy from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
copy to linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
index 3658919..24b14c8 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mm_malloc.h
@@ -10,6 +10,8 @@
 #ifndef _MM_MALLOC_H_INCLUDED
 #define _MM_MALLOC_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <stdlib.h>
 
 /* We can't depend on <stdlib.h> since the prototype of posix_memalign
@@ -25,12 +27,8 @@
 {
   /* PowerPC64 ELF V2 ABI requires quadword alignment.  */
   size_t vec_align = sizeof (__vector float);
-  /* Linux GLIBC malloc alignment is at least 2 X ptr size.  */
-  size_t malloc_align = (sizeof (void *) + sizeof (void *));
   void *ptr;
 
-  if (alignment == malloc_align && alignment == vec_align)
-    return malloc (size);
   if (alignment < vec_align)
     alignment = vec_align;
   if (posix_memalign (&ptr, alignment, size) == 0)
@@ -45,4 +43,8 @@
   free (ptr);
 }
 
+#else
+#include_next <mm_malloc.h>
+#endif
+
 #endif /* _MM_MALLOC_H_INCLUDED */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
index b949653..c55c447 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/mmintrin.h
@@ -35,6 +35,8 @@
 #ifndef _MMINTRIN_H_INCLUDED
 #define _MMINTRIN_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 #include <altivec.h>
 /* The Intel API is flexible enough that we must allow aliasing with other
    vector types, and their scalar components.  */
@@ -1440,4 +1442,9 @@
   return (res.as_m64);
 #endif
 }
+
+#else
+#include_next <mmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* _MMINTRIN_H_INCLUDED */
diff --git a/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h
new file mode 100644
index 0000000..6d93383
--- /dev/null
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/pmmintrin.h
@@ -0,0 +1,150 @@
+/*===---- pmmintrin.h - Implementation of SSE3 intrinsics on PowerPC -------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerpc64le.
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.
+
+   In the specific case of X86 SSE3 intrinsics, the PowerPC VMX/VSX ISA
+   is a good match for most SIMD operations.  However the Horizontal
+   add/sub requires the data pairs be permuted into a separate
+   registers with vertical even/odd alignment for the operation.
+   And the addsub operation requires the sign of only the even numbered
+   elements be flipped (xored with -0.0).
+   For larger blocks of code using these intrinsic implementations,
+   the compiler be should be able to schedule instructions to avoid
+   additional latency.
+
+   In the specific case of the monitor and mwait instructions there are
+   no direct equivalent in the PowerISA at this time.  So those
+   intrinsics are not implemented.  */
+#error "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this warning."
+#endif
+
+#ifndef PMMINTRIN_H_
+#define PMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+/* We need definitions from the SSE2 and SSE header files*/
+#include <emmintrin.h>
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_addsub_ps (__m128 __X, __m128 __Y)
+{
+  const __v4sf even_n0 = {-0.0, 0.0, -0.0, 0.0};
+  __v4sf even_neg_Y = vec_xor(__Y, even_n0);
+  return (__m128) vec_add (__X, even_neg_Y);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_addsub_pd (__m128d __X, __m128d __Y)
+{
+  const __v2df even_n0 = {-0.0, 0.0};
+  __v2df even_neg_Y = vec_xor(__Y, even_n0);
+  return (__m128d) vec_add (__X, even_neg_Y);
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_ps (__m128 __X, __m128 __Y)
+{
+  __vector unsigned char xform2 = {
+      0x00, 0x01, 0x02, 0x03,
+      0x08, 0x09, 0x0A, 0x0B,
+      0x10, 0x11, 0x12, 0x13,
+      0x18, 0x19, 0x1A, 0x1B
+    };
+  __vector unsigned char xform1 = {
+      0x04, 0x05, 0x06, 0x07,
+      0x0C, 0x0D, 0x0E, 0x0F,
+      0x14, 0x15, 0x16, 0x17,
+      0x1C, 0x1D, 0x1E, 0x1F
+    };
+  return (__m128) vec_add (vec_perm ((__v4sf) __X, (__v4sf) __Y, xform2),
+			   vec_perm ((__v4sf) __X, (__v4sf) __Y, xform1));
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_ps (__m128 __X, __m128 __Y)
+{
+  __vector unsigned char xform2 = {
+      0x00, 0x01, 0x02, 0x03,
+      0x08, 0x09, 0x0A, 0x0B,
+      0x10, 0x11, 0x12, 0x13,
+      0x18, 0x19, 0x1A, 0x1B
+    };
+  __vector unsigned char xform1 = {
+      0x04, 0x05, 0x06, 0x07,
+      0x0C, 0x0D, 0x0E, 0x0F,
+      0x14, 0x15, 0x16, 0x17,
+      0x1C, 0x1D, 0x1E, 0x1F
+    };
+  return (__m128) vec_sub (vec_perm ((__v4sf) __X, (__v4sf) __Y, xform2),
+			   vec_perm ((__v4sf) __X, (__v4sf) __Y, xform1));
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pd (__m128d __X, __m128d __Y)
+{
+  return (__m128d) vec_add (vec_mergeh ((__v2df) __X, (__v2df)__Y),
+				  vec_mergel ((__v2df) __X, (__v2df)__Y));
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pd (__m128d __X, __m128d __Y)
+{
+  return (__m128d) vec_sub (vec_mergeh ((__v2df) __X, (__v2df)__Y),
+			    vec_mergel ((__v2df) __X, (__v2df)__Y));
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_movehdup_ps (__m128 __X)
+{
+  return (__m128)vec_mergeo ((__v4su)__X, (__v4su)__X);
+}
+
+extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_moveldup_ps (__m128 __X)
+{
+  return (__m128)vec_mergee ((__v4su)__X, (__v4su)__X);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_loaddup_pd (double const *__P)
+{
+  return (__m128d) vec_splats (*__P);
+}
+
+extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_movedup_pd (__m128d __X)
+{
+  return _mm_shuffle_pd (__X, __X, _MM_SHUFFLE2 (0,0));
+}
+
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_lddqu_si128 (__m128i const *__P)
+{
+  return (__m128i) (vec_vsx_ld(0, (signed int const *)__P));
+}
+
+/* POWER8 / POWER9 have no equivalent for _mm_monitor nor _mm_wait.  */
+
+#else
+#include_next <pmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* PMMINTRIN_H_ */
diff --git a/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h
new file mode 100644
index 0000000..56ef6ba
--- /dev/null
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/smmintrin.h
@@ -0,0 +1,85 @@
+/*===---- smmintrin.h - Implementation of SSE4 intrinsics on PowerPC -------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.
+
+   NOTE: This is NOT a complete implementation of the SSE4 intrinsics!  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerp64/powerpc64le.
+
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.  */
+#error                                                                         \
+    "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this error."
+#endif
+
+#ifndef SMMINTRIN_H_
+#define SMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+#include <altivec.h>
+#include <emmintrin.h>
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi8(__m128i __X, const int __N) {
+  return (unsigned char)((__v16qi)__X)[__N & 15];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi32(__m128i __X, const int __N) {
+  return ((__v4si)__X)[__N & 3];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_epi64(__m128i __X, const int __N) {
+  return ((__v2di)__X)[__N & 1];
+}
+
+extern __inline int
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_extract_ps(__m128 __X, const int __N) {
+  return ((__v4si)__X)[__N & 3];
+}
+
+extern __inline __m128i
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_blend_epi16(__m128i __A, __m128i __B, const int __imm8) {
+  __v16qi __charmask = vec_splats((signed char)__imm8);
+  __charmask = vec_gb(__charmask);
+  __v8hu __shortmask = (__v8hu)vec_unpackh(__charmask);
+#ifdef __BIG_ENDIAN__
+  __shortmask = vec_reve(__shortmask);
+#endif
+  return (__m128i)vec_sel((__v8hu)__A, (__v8hu)__B, __shortmask);
+}
+
+extern __inline __m128i
+    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+    _mm_blendv_epi8(__m128i __A, __m128i __B, __m128i __mask) {
+  const __v16qu __seven = vec_splats((unsigned char)0x07);
+  __v16qu __lmask = vec_sra((__v16qu)__mask, __seven);
+  return (__m128i)vec_sel((__v16qu)__A, (__v16qu)__B, __lmask);
+}
+
+#else
+#include_next <smmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* _SMMINTRIN_H_ */
diff --git a/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h
new file mode 100644
index 0000000..b5a935d
--- /dev/null
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/tmmintrin.h
@@ -0,0 +1,495 @@
+/*===---- tmmintrin.h - Implementation of SSSE3 intrinsics on PowerPC ------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+/* Implemented from the specification included in the Intel C++ Compiler
+   User Guide and Reference, version 9.0.  */
+
+#ifndef NO_WARN_X86_INTRINSICS
+/* This header is distributed to simplify porting x86_64 code that
+   makes explicit use of Intel intrinsics to powerpc64le.
+
+   It is the user's responsibility to determine if the results are
+   acceptable and make additional changes as necessary.
+
+   Note that much code that uses Intel intrinsics can be rewritten in
+   standard C or GNU C extensions, which are more portable and better
+   optimized across multiple targets.  */
+#endif
+
+#ifndef TMMINTRIN_H_
+#define TMMINTRIN_H_
+
+#if defined(__linux__) && defined(__ppc64__)
+
+#include <altivec.h>
+
+/* We need definitions from the SSE header files.  */
+#include <pmmintrin.h>
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi16 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v8hi) __A);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi32 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v4si) __A);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_epi8 (__m128i __A)
+{
+  return (__m128i) vec_abs ((__v16qi) __A);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi16 (__m64 __A)
+{
+  __v8hi __B = (__v8hi) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi32 (__m64 __A)
+{
+  __v4si __B = (__v4si) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_abs_pi8 (__m64 __A)
+{
+  __v16qi __B = (__v16qi) (__v2du) { __A, __A };
+  return (__m64) ((__v2du) vec_abs (__B))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_alignr_epi8 (__m128i __A, __m128i __B, const unsigned int __count)
+{
+  if (__builtin_constant_p (__count) && __count < 16)
+    {
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_reve ((__v16qu) __A);
+      __B = (__m128i) vec_reve ((__v16qu) __B);
+#endif
+      __A = (__m128i) vec_sld ((__v16qu) __B, (__v16qu) __A, __count);
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_reve ((__v16qu) __A);
+#endif
+      return __A;
+    }
+
+  if (__count == 0)
+    return __B;
+
+  if (__count >= 16)
+    {
+      if (__count >= 32)
+	{
+	  const __v16qu zero = { 0 };
+	  return (__m128i) zero;
+	}
+      else
+	{
+	  const __v16qu __shift =
+	    vec_splats ((unsigned char) ((__count - 16) * 8));
+#ifdef __LITTLE_ENDIAN__
+	  return (__m128i) vec_sro ((__v16qu) __A, __shift);
+#else
+	  return (__m128i) vec_slo ((__v16qu) __A, __shift);
+#endif
+	}
+    }
+  else
+    {
+      const __v16qu __shiftA =
+	vec_splats ((unsigned char) ((16 - __count) * 8));
+      const __v16qu __shiftB = vec_splats ((unsigned char) (__count * 8));
+#ifdef __LITTLE_ENDIAN__
+      __A = (__m128i) vec_slo ((__v16qu) __A, __shiftA);
+      __B = (__m128i) vec_sro ((__v16qu) __B, __shiftB);
+#else
+      __A = (__m128i) vec_sro ((__v16qu) __A, __shiftA);
+      __B = (__m128i) vec_slo ((__v16qu) __B, __shiftB);
+#endif
+      return (__m128i) vec_or ((__v16qu) __A, (__v16qu) __B);
+    }
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_alignr_pi8 (__m64 __A, __m64 __B, unsigned int __count)
+{
+  if (__count < 16)
+    {
+      __v2du __C = { __B, __A };
+#ifdef __LITTLE_ENDIAN__
+      const __v4su __shift = { __count << 3, 0, 0, 0 };
+      __C = (__v2du) vec_sro ((__v16qu) __C, (__v16qu) __shift);
+#else
+      const __v4su __shift = { 0, 0, 0, __count << 3 };
+      __C = (__v2du) vec_slo ((__v16qu) __C, (__v16qu) __shift);
+#endif
+      return (__m64) __C[0];
+    }
+  else
+    {
+      const __m64 __zero = { 0 };
+      return __zero;
+    }
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_add (__C, __D);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_epi32 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 };
+  __v4si __C = vec_perm ((__v4si) __A, (__v4si) __B, __P);
+  __v4si __D = vec_perm ((__v4si) __A, (__v4si) __B, __Q);
+  return (__m128i) vec_add (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pi16 (__m64 __A, __m64 __B)
+{
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_add (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadd_pi32 (__m64 __A, __m64 __B)
+{
+  __v4si __C = (__v4si) (__v2du) { __A, __B };
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11,  0,  1,  2,  3,  8,  9, 10, 11 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15,  4,  5,  6,  7, 12, 13, 14, 15 };
+  __v4si __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_add (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadds_epi16 (__m128i __A, __m128i __B)
+{
+  __v4si __C = { 0 }, __D = { 0 };
+  __C = vec_sum4s ((__v8hi) __A, __C);
+  __D = vec_sum4s ((__v8hi) __B, __D);
+  __C = (__v4si) vec_packs (__C, __D);
+  return (__m128i) __C;
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hadds_pi16 (__m64 __A, __m64 __B)
+{
+  const __v4si __zero = { 0 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v4si __D = vec_sum4s (__C, __zero);
+  __C = vec_packs (__D, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_sub (__C, __D);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_epi32 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30, 31 };
+  __v4si __C = vec_perm ((__v4si) __A, (__v4si) __B, __P);
+  __v4si __D = vec_perm ((__v4si) __A, (__v4si) __B, __Q);
+  return (__m128i) vec_sub (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pi16 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v8hi __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_sub (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsub_pi32 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  2,  3,  8,  9, 10, 11,  0,  1,  2,  3,  8,  9, 10, 11 };
+  const __v16qu __Q =
+    {  4,  5,  6,  7, 12, 13, 14, 15,  4,  5,  6,  7, 12, 13, 14, 15 };
+  __v4si __C = (__v4si) (__v2du) { __A, __B };
+  __v4si __D = vec_perm (__C, __C, __Q);
+  __C = vec_perm (__C, __C, __P);
+  __C = vec_sub (__C, __D);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsubs_epi16 (__m128i __A, __m128i __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __v8hi __C = vec_perm ((__v8hi) __A, (__v8hi) __B, __P);
+  __v8hi __D = vec_perm ((__v8hi) __A, (__v8hi) __B, __Q);
+  return (__m128i) vec_subs (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_hsubs_pi16 (__m64 __A, __m64 __B)
+{
+  const __v16qu __P =
+    {  0,  1,  4,  5,  8,  9, 12, 13,  0,  1,  4,  5,  8,  9, 12, 13 };
+  const __v16qu __Q =
+    {  2,  3,  6,  7, 10, 11, 14, 15,  2,  3,  6,  7, 10, 11, 14, 15 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __B };
+  __v8hi __D = vec_perm (__C, __C, __P);
+  __v8hi __E = vec_perm (__C, __C, __Q);
+  __C = vec_subs (__D, __E);
+  return (__m64) ((__v2du) __C)[1];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_shuffle_epi8 (__m128i __A, __m128i __B)
+{
+  const __v16qi __zero = { 0 };
+  __vector __bool char __select = vec_cmplt ((__v16qi) __B, __zero);
+  __v16qi __C = vec_perm ((__v16qi) __A, (__v16qi) __A, (__v16qu) __B);
+  return (__m128i) vec_sel (__C, __zero, __select);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_shuffle_pi8 (__m64 __A, __m64 __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __C = (__v16qi) (__v2du) { __A, __A };
+  __v16qi __D = (__v16qi) (__v2du) { __B, __B };
+  __vector __bool char __select = vec_cmplt ((__v16qi) __D, __zero);
+  __C = vec_perm ((__v16qi) __C, (__v16qi) __C, (__v16qu) __D);
+  __C = vec_sel (__C, __zero, __select);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi8 (__m128i __A, __m128i __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __selectneg = (__v16qi) vec_cmplt ((__v16qi) __B, __zero);
+  __v16qi __selectpos =
+    (__v16qi) vec_neg ((__v16qi) vec_cmpgt ((__v16qi) __B, __zero));
+  __v16qi __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v16qi) __A, (__v16qi) __conv);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi16 (__m128i __A, __m128i __B)
+{
+  const __v8hi __zero = { 0 };
+  __v8hi __selectneg = (__v8hi) vec_cmplt ((__v8hi) __B, __zero);
+  __v8hi __selectpos =
+    (__v8hi) vec_neg ((__v8hi) vec_cmpgt ((__v8hi) __B, __zero));
+  __v8hi __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v8hi) __A, (__v8hi) __conv);
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_epi32 (__m128i __A, __m128i __B)
+{
+  const __v4si __zero = { 0 };
+  __v4si __selectneg = (__v4si) vec_cmplt ((__v4si) __B, __zero);
+  __v4si __selectpos =
+    (__v4si) vec_neg ((__v4si) vec_cmpgt ((__v4si) __B, __zero));
+  __v4si __conv = vec_add (__selectneg, __selectpos);
+  return (__m128i) vec_mul ((__v4si) __A, (__v4si) __conv);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi8 (__m64 __A, __m64 __B)
+{
+  const __v16qi __zero = { 0 };
+  __v16qi __C = (__v16qi) (__v2du) { __A, __A };
+  __v16qi __D = (__v16qi) (__v2du) { __B, __B };
+  __C = (__v16qi) _mm_sign_epi8 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi16 (__m64 __A, __m64 __B)
+{
+  const __v8hi __zero = { 0 };
+  __v8hi __C = (__v8hi) (__v2du) { __A, __A };
+  __v8hi __D = (__v8hi) (__v2du) { __B, __B };
+  __C = (__v8hi) _mm_sign_epi16 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_sign_pi32 (__m64 __A, __m64 __B)
+{
+  const __v4si __zero = { 0 };
+  __v4si __C = (__v4si) (__v2du) { __A, __A };
+  __v4si __D = (__v4si) (__v2du) { __B, __B };
+  __C = (__v4si) _mm_sign_epi32 ((__m128i) __C, (__m128i) __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maddubs_epi16 (__m128i __A, __m128i __B)
+{
+  __v8hi __unsigned = vec_splats ((signed short) 0x00ff);
+  __v8hi __C = vec_and (vec_unpackh ((__v16qi) __A), __unsigned);
+  __v8hi __D = vec_and (vec_unpackl ((__v16qi) __A), __unsigned);
+  __v8hi __E = vec_unpackh ((__v16qi) __B);
+  __v8hi __F = vec_unpackl ((__v16qi) __B);
+  __C = vec_mul (__C, __E);
+  __D = vec_mul (__D, __F);
+  const __v16qu __odds  =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __evens =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __E = vec_perm (__C, __D, __odds);
+  __F = vec_perm (__C, __D, __evens);
+  return (__m128i) vec_adds (__E, __F);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maddubs_pi16 (__m64 __A, __m64 __B)
+{
+  __v8hi __C = (__v8hi) (__v2du) { __A, __A };
+  __C = vec_unpackl ((__v16qi) __C);
+  const __v8hi __unsigned = vec_splats ((signed short) 0x00ff);
+  __C = vec_and (__C, __unsigned);
+  __v8hi __D = (__v8hi) (__v2du) { __B, __B };
+  __D = vec_unpackl ((__v16qi) __D);
+  __D = vec_mul (__C, __D);
+  const __v16qu __odds  =
+    {  0,  1,  4,  5,  8,  9, 12, 13, 16, 17, 20, 21, 24, 25, 28, 29 };
+  const __v16qu __evens =
+    {  2,  3,  6,  7, 10, 11, 14, 15, 18, 19, 22, 23, 26, 27, 30, 31 };
+  __C = vec_perm (__D, __D, __odds);
+  __D = vec_perm (__D, __D, __evens);
+  __C = vec_adds (__C, __D);
+  return (__m64) ((__v2du) (__C))[0];
+}
+
+extern __inline __m128i
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mulhrs_epi16 (__m128i __A, __m128i __B)
+{
+  __v4si __C = vec_unpackh ((__v8hi) __A);
+  __v4si __D = vec_unpackh ((__v8hi) __B);
+  __C = vec_mul (__C, __D);
+  __D = vec_unpackl ((__v8hi) __A);
+  __v4si __E = vec_unpackl ((__v8hi) __B);
+  __D = vec_mul (__D, __E);
+  const __v4su __shift = vec_splats ((unsigned int) 14);
+  __C = vec_sr (__C, __shift);
+  __D = vec_sr (__D, __shift);
+  const __v4si __ones = vec_splats ((signed int) 1);
+  __C = vec_add (__C, __ones);
+  __C = vec_sr (__C, (__v4su) __ones);
+  __D = vec_add (__D, __ones);
+  __D = vec_sr (__D, (__v4su) __ones);
+  return (__m128i) vec_pack (__C, __D);
+}
+
+extern __inline __m64
+__attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mulhrs_pi16 (__m64 __A, __m64 __B)
+{
+  __v4si __C = (__v4si) (__v2du) { __A, __A };
+  __C = vec_unpackh ((__v8hi) __C);
+  __v4si __D = (__v4si) (__v2du) { __B, __B };
+  __D = vec_unpackh ((__v8hi) __D);
+  __C = vec_mul (__C, __D);
+  const __v4su __shift = vec_splats ((unsigned int) 14);
+  __C = vec_sr (__C, __shift);
+  const __v4si __ones = vec_splats ((signed int) 1);
+  __C = vec_add (__C, __ones);
+  __C = vec_sr (__C, (__v4su) __ones);
+  __v8hi __E = vec_pack (__C, __D);
+  return (__m64) ((__v2du) (__E))[0];
+}
+
+#else
+#include_next <tmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
+#endif /* TMMINTRIN_H_ */
diff --git a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
index 1b322b6..0f429fa 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/ppc_wrappers/xmmintrin.h
@@ -34,6 +34,8 @@
 #ifndef _XMMINTRIN_H_INCLUDED
 #define _XMMINTRIN_H_INCLUDED
 
+#if defined(__linux__) && defined(__ppc64__)
+
 /* Define four value permute mask */
 #define _MM_SHUFFLE(w,x,y,z) (((w) << 6) | ((x) << 4) | ((y) << 2) | (z))
 
@@ -1835,4 +1837,8 @@
 /* For backward source compatibility.  */
 //# include <emmintrin.h>
 
+#else
+#include_next <xmmintrin.h>
+#endif /* defined(__linux__) && defined(__ppc64__) */
+
 #endif /* _XMMINTRIN_H_INCLUDED */
diff --git a/linux-x86/lib64/clang/9.0.7/include/prfchwintrin.h b/linux-x86/lib64/clang/10.0.1/include/prfchwintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/prfchwintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/prfchwintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/ptwriteintrin.h b/linux-x86/lib64/clang/10.0.1/include/ptwriteintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/ptwriteintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/ptwriteintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/rdseedintrin.h b/linux-x86/lib64/clang/10.0.1/include/rdseedintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/rdseedintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/rdseedintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/rtmintrin.h b/linux-x86/lib64/clang/10.0.1/include/rtmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/rtmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/rtmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/s390intrin.h b/linux-x86/lib64/clang/10.0.1/include/s390intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/s390intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/s390intrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/allocator_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/allocator_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/allocator_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/allocator_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/asan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/asan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/asan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/asan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/common_interface_defs.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/common_interface_defs.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/common_interface_defs.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/common_interface_defs.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/coverage_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/coverage_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/coverage_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/coverage_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/dfsan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/dfsan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/dfsan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/dfsan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/hwasan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/hwasan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/hwasan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/hwasan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/linux_syscall_hooks.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/linux_syscall_hooks.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/linux_syscall_hooks.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/linux_syscall_hooks.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/lsan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/lsan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/lsan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/lsan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/msan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/msan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/msan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/msan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/netbsd_syscall_hooks.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/netbsd_syscall_hooks.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/netbsd_syscall_hooks.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/netbsd_syscall_hooks.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/scudo_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/scudo_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/scudo_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/scudo_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface_atomic.h b/linux-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface_atomic.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sanitizer/tsan_interface_atomic.h
rename to linux-x86/lib64/clang/10.0.1/include/sanitizer/tsan_interface_atomic.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/sgxintrin.h b/linux-x86/lib64/clang/10.0.1/include/sgxintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/sgxintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/sgxintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/shaintrin.h b/linux-x86/lib64/clang/10.0.1/include/shaintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/shaintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/shaintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/smmintrin.h b/linux-x86/lib64/clang/10.0.1/include/smmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/smmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/smmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdalign.h b/linux-x86/lib64/clang/10.0.1/include/stdalign.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdalign.h
rename to linux-x86/lib64/clang/10.0.1/include/stdalign.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdarg.h b/linux-x86/lib64/clang/10.0.1/include/stdarg.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdarg.h
rename to linux-x86/lib64/clang/10.0.1/include/stdarg.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdatomic.h b/linux-x86/lib64/clang/10.0.1/include/stdatomic.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdatomic.h
rename to linux-x86/lib64/clang/10.0.1/include/stdatomic.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdbool.h b/linux-x86/lib64/clang/10.0.1/include/stdbool.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdbool.h
rename to linux-x86/lib64/clang/10.0.1/include/stdbool.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stddef.h b/linux-x86/lib64/clang/10.0.1/include/stddef.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stddef.h
rename to linux-x86/lib64/clang/10.0.1/include/stddef.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdint.h b/linux-x86/lib64/clang/10.0.1/include/stdint.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdint.h
rename to linux-x86/lib64/clang/10.0.1/include/stdint.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/stdnoreturn.h b/linux-x86/lib64/clang/10.0.1/include/stdnoreturn.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/stdnoreturn.h
rename to linux-x86/lib64/clang/10.0.1/include/stdnoreturn.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/tbmintrin.h b/linux-x86/lib64/clang/10.0.1/include/tbmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/tbmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/tbmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/tgmath.h b/linux-x86/lib64/clang/10.0.1/include/tgmath.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/tgmath.h
rename to linux-x86/lib64/clang/10.0.1/include/tgmath.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/tmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/tmmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/tmmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/tmmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/unwind.h b/linux-x86/lib64/clang/10.0.1/include/unwind.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/unwind.h
rename to linux-x86/lib64/clang/10.0.1/include/unwind.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/vadefs.h b/linux-x86/lib64/clang/10.0.1/include/vadefs.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/vadefs.h
rename to linux-x86/lib64/clang/10.0.1/include/vadefs.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/vaesintrin.h b/linux-x86/lib64/clang/10.0.1/include/vaesintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/vaesintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/vaesintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/varargs.h b/linux-x86/lib64/clang/10.0.1/include/varargs.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/varargs.h
rename to linux-x86/lib64/clang/10.0.1/include/varargs.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/vecintrin.h b/linux-x86/lib64/clang/10.0.1/include/vecintrin.h
similarity index 95%
copy from darwin-x86/lib64/clang/9.0.7/include/vecintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/vecintrin.h
index 115d893..c71b76a 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/vecintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/vecintrin.h
@@ -517,6 +517,141 @@
 }
 #endif
 
+/*-- vec_revb ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_revb(vector signed short __vec) {
+  return (vector signed short)
+         __builtin_s390_vlbrh((vector unsigned short)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_revb(vector unsigned short __vec) {
+  return __builtin_s390_vlbrh(__vec);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_revb(vector signed int __vec) {
+  return (vector signed int)
+         __builtin_s390_vlbrf((vector unsigned int)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_revb(vector unsigned int __vec) {
+  return __builtin_s390_vlbrf(__vec);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_revb(vector signed long long __vec) {
+  return (vector signed long long)
+         __builtin_s390_vlbrg((vector unsigned long long)__vec);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_revb(vector unsigned long long __vec) {
+  return __builtin_s390_vlbrg(__vec);
+}
+
+#if __ARCH__ >= 12
+static inline __ATTRS_o_ai vector float
+vec_revb(vector float __vec) {
+  return (vector float)
+         __builtin_s390_vlbrf((vector unsigned int)__vec);
+}
+#endif
+
+static inline __ATTRS_o_ai vector double
+vec_revb(vector double __vec) {
+  return (vector double)
+         __builtin_s390_vlbrg((vector unsigned long long)__vec);
+}
+
+/*-- vec_reve ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_reve(vector signed char __vec) {
+  return (vector signed char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                                __vec[11], __vec[10], __vec[9], __vec[8],
+                                __vec[7], __vec[6], __vec[5], __vec[4],
+                                __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_reve(vector unsigned char __vec) {
+  return (vector unsigned char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                                  __vec[11], __vec[10], __vec[9], __vec[8],
+                                  __vec[7], __vec[6], __vec[5], __vec[4],
+                                  __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_reve(vector bool char __vec) {
+  return (vector bool char) { __vec[15], __vec[14], __vec[13], __vec[12],
+                              __vec[11], __vec[10], __vec[9], __vec[8],
+                              __vec[7], __vec[6], __vec[5], __vec[4],
+                              __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_reve(vector signed short __vec) {
+  return (vector signed short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                                 __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_reve(vector unsigned short __vec) {
+  return (vector unsigned short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                                   __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_reve(vector bool short __vec) {
+  return (vector bool short) { __vec[7], __vec[6], __vec[5], __vec[4],
+                               __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_reve(vector signed int __vec) {
+  return (vector signed int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_reve(vector unsigned int __vec) {
+  return (vector unsigned int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_reve(vector bool int __vec) {
+  return (vector bool int) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_reve(vector signed long long __vec) {
+  return (vector signed long long) { __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_reve(vector unsigned long long __vec) {
+  return (vector unsigned long long) { __vec[1], __vec[0] };
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_reve(vector bool long long __vec) {
+  return (vector bool long long) { __vec[1], __vec[0] };
+}
+
+#if __ARCH__ >= 12
+static inline __ATTRS_o_ai vector float
+vec_reve(vector float __vec) {
+  return (vector float) { __vec[3], __vec[2], __vec[1], __vec[0] };
+}
+#endif
+
+static inline __ATTRS_o_ai vector double
+vec_reve(vector double __vec) {
+  return (vector double) { __vec[1], __vec[0] };
+}
+
 /*-- vec_sel ----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -6835,6 +6970,56 @@
   __builtin_s390_vsldb((vector unsigned char)(X), \
                        (vector unsigned char)(Y), (Z) * 4))
 
+/*-- vec_sldb ---------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+extern __ATTRS_o vector signed char
+vec_sldb(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned char
+vec_sldb(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed short
+vec_sldb(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned short
+vec_sldb(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed int
+vec_sldb(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned int
+vec_sldb(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed long long
+vec_sldb(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned long long
+vec_sldb(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector float
+vec_sldb(vector float __a, vector float __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector double
+vec_sldb(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 7);
+
+#define vec_sldb(X, Y, Z) ((__typeof__((vec_sldb)((X), (Y), (Z)))) \
+  __builtin_s390_vsld((vector unsigned char)(X), \
+                      (vector unsigned char)(Y), (Z)))
+
+#endif
+
 /*-- vec_sral ---------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -7565,6 +7750,56 @@
     (vector unsigned char)__a, (vector unsigned char)__b);
 }
 
+/*-- vec_srdb ---------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+extern __ATTRS_o vector signed char
+vec_srdb(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned char
+vec_srdb(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed short
+vec_srdb(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned short
+vec_srdb(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed int
+vec_srdb(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned int
+vec_srdb(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector signed long long
+vec_srdb(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector unsigned long long
+vec_srdb(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector float
+vec_srdb(vector float __a, vector float __b, int __c)
+  __constant_range(__c, 0, 7);
+
+extern __ATTRS_o vector double
+vec_srdb(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 7);
+
+#define vec_srdb(X, Y, Z) ((__typeof__((vec_srdb)((X), (Y), (Z)))) \
+  __builtin_s390_vsrd((vector unsigned char)(X), \
+                      (vector unsigned char)(Y), (Z)))
+
+#endif
+
 /*-- vec_abs ----------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed char
@@ -8711,6 +8946,22 @@
   return __builtin_convertvector(__a, vector double);
 }
 
+/*-- vec_float --------------------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector float
+vec_float(vector signed int __a) {
+  return __builtin_convertvector(__a, vector float);
+}
+
+static inline __ATTRS_o_ai vector float
+vec_float(vector unsigned int __a) {
+  return __builtin_convertvector(__a, vector float);
+}
+
+#endif
+
 /*-- vec_signed -------------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector signed long long
@@ -8718,6 +8969,13 @@
   return __builtin_convertvector(__a, vector signed long long);
 }
 
+#if __ARCH__ >= 13
+static inline __ATTRS_o_ai vector signed int
+vec_signed(vector float __a) {
+  return __builtin_convertvector(__a, vector signed int);
+}
+#endif
+
 /*-- vec_unsigned -----------------------------------------------------------*/
 
 static inline __ATTRS_o_ai vector unsigned long long
@@ -8725,6 +8983,13 @@
   return __builtin_convertvector(__a, vector unsigned long long);
 }
 
+#if __ARCH__ >= 13
+static inline __ATTRS_o_ai vector unsigned int
+vec_unsigned(vector float __a) {
+  return __builtin_convertvector(__a, vector unsigned int);
+}
+#endif
+
 /*-- vec_roundp -------------------------------------------------------------*/
 
 #if __ARCH__ >= 12
@@ -10442,6 +10707,147 @@
   return __builtin_s390_vfaezfs(__a, __b, 8, __cc);
 }
 
+/*-- vec_search_string_cc ---------------------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed char __a, vector signed char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb((vector unsigned char)__a,
+                               (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool char __a, vector bool char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb((vector unsigned char)__a,
+                               (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned char __a, vector unsigned char __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsb(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed short __a, vector signed short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh((vector unsigned short)__a,
+                               (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool short __a, vector bool short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh((vector unsigned short)__a,
+                               (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned short __a, vector unsigned short __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsh(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector signed int __a, vector signed int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf((vector unsigned int)__a,
+                               (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector bool int __a, vector bool int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf((vector unsigned int)__a,
+                               (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_cc(vector unsigned int __a, vector unsigned int __b,
+                     vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrsf(__a, __b, __c, __cc);
+}
+
+#endif
+
+/*-- vec_search_string_until_zero_cc ----------------------------------------*/
+
+#if __ARCH__ >= 13
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed char __a,
+                                vector signed char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb((vector unsigned char)__a,
+                                (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool char __a,
+                                vector bool char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb((vector unsigned char)__a,
+                                (vector unsigned char)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned char __a,
+                                vector unsigned char __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszb(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed short __a,
+                                vector signed short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh((vector unsigned short)__a,
+                                (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool short __a,
+                                vector bool short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh((vector unsigned short)__a,
+                                (vector unsigned short)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned short __a,
+                                vector unsigned short __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszh(__a, __b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector signed int __a,
+                                vector signed int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf((vector unsigned int)__a,
+                                (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector bool int __a,
+                                vector bool int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf((vector unsigned int)__a,
+                                (vector unsigned int)__b, __c, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_search_string_until_zero_cc(vector unsigned int __a,
+                                vector unsigned int __b,
+                                vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrszf(__a, __b, __c, __cc);
+}
+
+#endif
+
 #undef __constant_pow2_range
 #undef __constant_range
 #undef __constant
diff --git a/linux-x86/lib64/clang/9.0.7/include/vpclmulqdqintrin.h b/linux-x86/lib64/clang/10.0.1/include/vpclmulqdqintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/vpclmulqdqintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/vpclmulqdqintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/waitpkgintrin.h b/linux-x86/lib64/clang/10.0.1/include/waitpkgintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/waitpkgintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/waitpkgintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/wbnoinvdintrin.h b/linux-x86/lib64/clang/10.0.1/include/wbnoinvdintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/wbnoinvdintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/wbnoinvdintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/wmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/wmmintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/wmmintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/wmmintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/x86intrin.h b/linux-x86/lib64/clang/10.0.1/include/x86intrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/x86intrin.h
rename to linux-x86/lib64/clang/10.0.1/include/x86intrin.h
diff --git a/darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h b/linux-x86/lib64/clang/10.0.1/include/xmmintrin.h
similarity index 99%
copy from darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h
copy to linux-x86/lib64/clang/10.0.1/include/xmmintrin.h
index 6f5517e..75ff376 100644
--- a/darwin-x86/lib64/clang/9.0.7/include/xmmintrin.h
+++ b/linux-x86/lib64/clang/10.0.1/include/xmmintrin.h
@@ -1919,7 +1919,11 @@
 static __inline__ void __DEFAULT_FN_ATTRS
 _mm_storeh_pi(__m64 *__p, __m128 __a)
 {
-  __builtin_ia32_storehps((__v2si *)__p, (__v4sf)__a);
+  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_storeh_pi_struct {
+    __mm_storeh_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 2, 3);
 }
 
 /// Stores the lower 64 bits of a 128-bit vector of [4 x float] to a
@@ -1936,7 +1940,11 @@
 static __inline__ void __DEFAULT_FN_ATTRS
 _mm_storel_pi(__m64 *__p, __m128 __a)
 {
-  __builtin_ia32_storelps((__v2si *)__p, (__v4sf)__a);
+  typedef float __mm_storeh_pi_v2f32 __attribute__((__vector_size__(8)));
+  struct __mm_storeh_pi_struct {
+    __mm_storeh_pi_v2f32 __u;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __mm_storeh_pi_struct*)__p)->__u = __builtin_shufflevector(__a, __a, 0, 1);
 }
 
 /// Stores the lower 32 bits of a 128-bit vector of [4 x float] to a
diff --git a/linux-x86/lib64/clang/9.0.7/include/xopintrin.h b/linux-x86/lib64/clang/10.0.1/include/xopintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xopintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xopintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/xsavecintrin.h b/linux-x86/lib64/clang/10.0.1/include/xsavecintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xsavecintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xsavecintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/xsaveintrin.h b/linux-x86/lib64/clang/10.0.1/include/xsaveintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xsaveintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xsaveintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/xsaveoptintrin.h b/linux-x86/lib64/clang/10.0.1/include/xsaveoptintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xsaveoptintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xsaveoptintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/xsavesintrin.h b/linux-x86/lib64/clang/10.0.1/include/xsavesintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xsavesintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xsavesintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/include/xtestintrin.h b/linux-x86/lib64/clang/10.0.1/include/xtestintrin.h
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/include/xtestintrin.h
rename to linux-x86/lib64/clang/10.0.1/include/xtestintrin.h
diff --git a/linux-x86/lib64/clang/9.0.7/share/asan_blacklist.txt b/linux-x86/lib64/clang/10.0.1/share/asan_blacklist.txt
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/share/asan_blacklist.txt
rename to linux-x86/lib64/clang/10.0.1/share/asan_blacklist.txt
diff --git a/linux-x86/lib64/clang/9.0.7/share/cfi_blacklist.txt b/linux-x86/lib64/clang/10.0.1/share/cfi_blacklist.txt
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/share/cfi_blacklist.txt
rename to linux-x86/lib64/clang/10.0.1/share/cfi_blacklist.txt
diff --git a/linux-x86/lib64/clang/9.0.7/share/dfsan_abilist.txt b/linux-x86/lib64/clang/10.0.1/share/dfsan_abilist.txt
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/share/dfsan_abilist.txt
rename to linux-x86/lib64/clang/10.0.1/share/dfsan_abilist.txt
diff --git a/linux-x86/lib64/clang/9.0.7/share/hwasan_blacklist.txt b/linux-x86/lib64/clang/10.0.1/share/hwasan_blacklist.txt
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/share/hwasan_blacklist.txt
rename to linux-x86/lib64/clang/10.0.1/share/hwasan_blacklist.txt
diff --git a/linux-x86/lib64/clang/9.0.7/share/msan_blacklist.txt b/linux-x86/lib64/clang/10.0.1/share/msan_blacklist.txt
similarity index 100%
rename from linux-x86/lib64/clang/9.0.7/share/msan_blacklist.txt
rename to linux-x86/lib64/clang/10.0.1/share/msan_blacklist.txt
diff --git a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h b/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
deleted file mode 100644
index 2970d17..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/__clang_cuda_intrinsics.h
+++ /dev/null
@@ -1,478 +0,0 @@
-/*===--- __clang_cuda_intrinsics.h - Device-side CUDA intrinsic wrappers ---===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-#ifndef __CLANG_CUDA_INTRINSICS_H__
-#define __CLANG_CUDA_INTRINSICS_H__
-#ifndef __CUDA__
-#error "This file is for CUDA compilation only."
-#endif
-
-// sm_30 intrinsics: __shfl_{up,down,xor}.
-
-#define __SM_30_INTRINSICS_H__
-#define __SM_30_INTRINSICS_HPP__
-
-#if !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 300
-
-#pragma push_macro("__MAKE_SHUFFLES")
-#define __MAKE_SHUFFLES(__FnName, __IntIntrinsic, __FloatIntrinsic, __Mask,    \
-                        __Type)                                                \
-  inline __device__ int __FnName(int __val, __Type __offset,                   \
-                                 int __width = warpSize) {                     \
-    return __IntIntrinsic(__val, __offset,                                     \
-                          ((warpSize - __width) << 8) | (__Mask));             \
-  }                                                                            \
-  inline __device__ float __FnName(float __val, __Type __offset,               \
-                                   int __width = warpSize) {                   \
-    return __FloatIntrinsic(__val, __offset,                                   \
-                            ((warpSize - __width) << 8) | (__Mask));           \
-  }                                                                            \
-  inline __device__ unsigned int __FnName(unsigned int __val, __Type __offset, \
-                                          int __width = warpSize) {            \
-    return static_cast<unsigned int>(                                          \
-        ::__FnName(static_cast<int>(__val), __offset, __width));               \
-  }                                                                            \
-  inline __device__ long long __FnName(long long __val, __Type __offset,       \
-                                       int __width = warpSize) {               \
-    struct __Bits {                                                            \
-      int __a, __b;                                                            \
-    };                                                                         \
-    _Static_assert(sizeof(__val) == sizeof(__Bits));                           \
-    _Static_assert(sizeof(__Bits) == 2 * sizeof(int));                         \
-    __Bits __tmp;                                                              \
-    memcpy(&__val, &__tmp, sizeof(__val));                                     \
-    __tmp.__a = ::__FnName(__tmp.__a, __offset, __width);                      \
-    __tmp.__b = ::__FnName(__tmp.__b, __offset, __width);                      \
-    long long __ret;                                                           \
-    memcpy(&__ret, &__tmp, sizeof(__tmp));                                     \
-    return __ret;                                                              \
-  }                                                                            \
-  inline __device__ long __FnName(long __val, __Type __offset,                 \
-                                  int __width = warpSize) {                    \
-    _Static_assert(sizeof(long) == sizeof(long long) ||                        \
-                   sizeof(long) == sizeof(int));                               \
-    if (sizeof(long) == sizeof(long long)) {                                   \
-      return static_cast<long>(                                                \
-          ::__FnName(static_cast<long long>(__val), __offset, __width));       \
-    } else if (sizeof(long) == sizeof(int)) {                                  \
-      return static_cast<long>(                                                \
-          ::__FnName(static_cast<int>(__val), __offset, __width));             \
-    }                                                                          \
-  }                                                                            \
-  inline __device__ unsigned long __FnName(                                    \
-      unsigned long __val, __Type __offset, int __width = warpSize) {          \
-    return static_cast<unsigned long>(                                         \
-        ::__FnName(static_cast<long>(__val), __offset, __width));              \
-  }                                                                            \
-  inline __device__ unsigned long long __FnName(                               \
-      unsigned long long __val, __Type __offset, int __width = warpSize) {     \
-    return static_cast<unsigned long long>(::__FnName(                         \
-        static_cast<unsigned long long>(__val), __offset, __width));           \
-  }                                                                            \
-  inline __device__ double __FnName(double __val, __Type __offset,             \
-                                    int __width = warpSize) {                  \
-    long long __tmp;                                                           \
-    _Static_assert(sizeof(__tmp) == sizeof(__val));                            \
-    memcpy(&__tmp, &__val, sizeof(__val));                                     \
-    __tmp = ::__FnName(__tmp, __offset, __width);                              \
-    double __ret;                                                              \
-    memcpy(&__ret, &__tmp, sizeof(__ret));                                     \
-    return __ret;                                                              \
-  }
-
-__MAKE_SHUFFLES(__shfl, __nvvm_shfl_idx_i32, __nvvm_shfl_idx_f32, 0x1f, int);
-// We use 0 rather than 31 as our mask, because shfl.up applies to lanes >=
-// maxLane.
-__MAKE_SHUFFLES(__shfl_up, __nvvm_shfl_up_i32, __nvvm_shfl_up_f32, 0,
-                unsigned int);
-__MAKE_SHUFFLES(__shfl_down, __nvvm_shfl_down_i32, __nvvm_shfl_down_f32, 0x1f,
-                unsigned int);
-__MAKE_SHUFFLES(__shfl_xor, __nvvm_shfl_bfly_i32, __nvvm_shfl_bfly_f32, 0x1f,
-                int);
-#pragma pop_macro("__MAKE_SHUFFLES")
-
-#endif // !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 300
-
-#if CUDA_VERSION >= 9000
-#if (!defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 300)
-// __shfl_sync_* variants available in CUDA-9
-#pragma push_macro("__MAKE_SYNC_SHUFFLES")
-#define __MAKE_SYNC_SHUFFLES(__FnName, __IntIntrinsic, __FloatIntrinsic,       \
-                             __Mask, __Type)                                   \
-  inline __device__ int __FnName(unsigned int __mask, int __val,               \
-                                 __Type __offset, int __width = warpSize) {    \
-    return __IntIntrinsic(__mask, __val, __offset,                             \
-                          ((warpSize - __width) << 8) | (__Mask));             \
-  }                                                                            \
-  inline __device__ float __FnName(unsigned int __mask, float __val,           \
-                                   __Type __offset, int __width = warpSize) {  \
-    return __FloatIntrinsic(__mask, __val, __offset,                           \
-                            ((warpSize - __width) << 8) | (__Mask));           \
-  }                                                                            \
-  inline __device__ unsigned int __FnName(unsigned int __mask,                 \
-                                          unsigned int __val, __Type __offset, \
-                                          int __width = warpSize) {            \
-    return static_cast<unsigned int>(                                          \
-        ::__FnName(__mask, static_cast<int>(__val), __offset, __width));       \
-  }                                                                            \
-  inline __device__ long long __FnName(unsigned int __mask, long long __val,   \
-                                       __Type __offset,                        \
-                                       int __width = warpSize) {               \
-    struct __Bits {                                                            \
-      int __a, __b;                                                            \
-    };                                                                         \
-    _Static_assert(sizeof(__val) == sizeof(__Bits));                           \
-    _Static_assert(sizeof(__Bits) == 2 * sizeof(int));                         \
-    __Bits __tmp;                                                              \
-    memcpy(&__val, &__tmp, sizeof(__val));                                     \
-    __tmp.__a = ::__FnName(__mask, __tmp.__a, __offset, __width);              \
-    __tmp.__b = ::__FnName(__mask, __tmp.__b, __offset, __width);              \
-    long long __ret;                                                           \
-    memcpy(&__ret, &__tmp, sizeof(__tmp));                                     \
-    return __ret;                                                              \
-  }                                                                            \
-  inline __device__ unsigned long long __FnName(                               \
-      unsigned int __mask, unsigned long long __val, __Type __offset,          \
-      int __width = warpSize) {                                                \
-    return static_cast<unsigned long long>(::__FnName(                         \
-        __mask, static_cast<unsigned long long>(__val), __offset, __width));   \
-  }                                                                            \
-  inline __device__ long __FnName(unsigned int __mask, long __val,             \
-                                  __Type __offset, int __width = warpSize) {   \
-    _Static_assert(sizeof(long) == sizeof(long long) ||                        \
-                   sizeof(long) == sizeof(int));                               \
-    if (sizeof(long) == sizeof(long long)) {                                   \
-      return static_cast<long>(::__FnName(                                     \
-          __mask, static_cast<long long>(__val), __offset, __width));          \
-    } else if (sizeof(long) == sizeof(int)) {                                  \
-      return static_cast<long>(                                                \
-          ::__FnName(__mask, static_cast<int>(__val), __offset, __width));     \
-    }                                                                          \
-  }                                                                            \
-  inline __device__ unsigned long __FnName(                                    \
-      unsigned int __mask, unsigned long __val, __Type __offset,               \
-      int __width = warpSize) {                                                \
-    return static_cast<unsigned long>(                                         \
-        ::__FnName(__mask, static_cast<long>(__val), __offset, __width));      \
-  }                                                                            \
-  inline __device__ double __FnName(unsigned int __mask, double __val,         \
-                                    __Type __offset, int __width = warpSize) { \
-    long long __tmp;                                                           \
-    _Static_assert(sizeof(__tmp) == sizeof(__val));                            \
-    memcpy(&__tmp, &__val, sizeof(__val));                                     \
-    __tmp = ::__FnName(__mask, __tmp, __offset, __width);                      \
-    double __ret;                                                              \
-    memcpy(&__ret, &__tmp, sizeof(__ret));                                     \
-    return __ret;                                                              \
-  }
-__MAKE_SYNC_SHUFFLES(__shfl_sync, __nvvm_shfl_sync_idx_i32,
-                     __nvvm_shfl_sync_idx_f32, 0x1f, int);
-// We use 0 rather than 31 as our mask, because shfl.up applies to lanes >=
-// maxLane.
-__MAKE_SYNC_SHUFFLES(__shfl_up_sync, __nvvm_shfl_sync_up_i32,
-                     __nvvm_shfl_sync_up_f32, 0, unsigned int);
-__MAKE_SYNC_SHUFFLES(__shfl_down_sync, __nvvm_shfl_sync_down_i32,
-                     __nvvm_shfl_sync_down_f32, 0x1f, unsigned int);
-__MAKE_SYNC_SHUFFLES(__shfl_xor_sync, __nvvm_shfl_sync_bfly_i32,
-                     __nvvm_shfl_sync_bfly_f32, 0x1f, int);
-#pragma pop_macro("__MAKE_SYNC_SHUFFLES")
-
-inline __device__ void __syncwarp(unsigned int mask = 0xffffffff) {
-  return __nvvm_bar_warp_sync(mask);
-}
-
-inline __device__ void __barrier_sync(unsigned int id) {
-  __nvvm_barrier_sync(id);
-}
-
-inline __device__ void __barrier_sync_count(unsigned int id,
-                                            unsigned int count) {
-  __nvvm_barrier_sync_cnt(id, count);
-}
-
-inline __device__ int __all_sync(unsigned int mask, int pred) {
-  return __nvvm_vote_all_sync(mask, pred);
-}
-
-inline __device__ int __any_sync(unsigned int mask, int pred) {
-  return __nvvm_vote_any_sync(mask, pred);
-}
-
-inline __device__ int __uni_sync(unsigned int mask, int pred) {
-  return __nvvm_vote_uni_sync(mask, pred);
-}
-
-inline __device__ unsigned int __ballot_sync(unsigned int mask, int pred) {
-  return __nvvm_vote_ballot_sync(mask, pred);
-}
-
-inline __device__ unsigned int __activemask() { return __nvvm_vote_ballot(1); }
-
-inline __device__ unsigned int __fns(unsigned mask, unsigned base, int offset) {
-  return __nvvm_fns(mask, base, offset);
-}
-
-#endif // !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 300
-
-// Define __match* builtins CUDA-9 headers expect to see.
-#if !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 700
-inline __device__ unsigned int __match32_any_sync(unsigned int mask,
-                                                  unsigned int value) {
-  return __nvvm_match_any_sync_i32(mask, value);
-}
-
-inline __device__ unsigned long long
-__match64_any_sync(unsigned int mask, unsigned long long value) {
-  return __nvvm_match_any_sync_i64(mask, value);
-}
-
-inline __device__ unsigned int
-__match32_all_sync(unsigned int mask, unsigned int value, int *pred) {
-  return __nvvm_match_all_sync_i32p(mask, value, pred);
-}
-
-inline __device__ unsigned long long
-__match64_all_sync(unsigned int mask, unsigned long long value, int *pred) {
-  return __nvvm_match_all_sync_i64p(mask, value, pred);
-}
-#include "crt/sm_70_rt.hpp"
-
-#endif // !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 700
-#endif // __CUDA_VERSION >= 9000
-
-// sm_32 intrinsics: __ldg and __funnelshift_{l,lc,r,rc}.
-
-// Prevent the vanilla sm_32 intrinsics header from being included.
-#define __SM_32_INTRINSICS_H__
-#define __SM_32_INTRINSICS_HPP__
-
-#if !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 320
-
-inline __device__ char __ldg(const char *ptr) { return __nvvm_ldg_c(ptr); }
-inline __device__ short __ldg(const short *ptr) { return __nvvm_ldg_s(ptr); }
-inline __device__ int __ldg(const int *ptr) { return __nvvm_ldg_i(ptr); }
-inline __device__ long __ldg(const long *ptr) { return __nvvm_ldg_l(ptr); }
-inline __device__ long long __ldg(const long long *ptr) {
-  return __nvvm_ldg_ll(ptr);
-}
-inline __device__ unsigned char __ldg(const unsigned char *ptr) {
-  return __nvvm_ldg_uc(ptr);
-}
-inline __device__ signed char __ldg(const signed char *ptr) {
-  return __nvvm_ldg_uc((const unsigned char *)ptr);
-}
-inline __device__ unsigned short __ldg(const unsigned short *ptr) {
-  return __nvvm_ldg_us(ptr);
-}
-inline __device__ unsigned int __ldg(const unsigned int *ptr) {
-  return __nvvm_ldg_ui(ptr);
-}
-inline __device__ unsigned long __ldg(const unsigned long *ptr) {
-  return __nvvm_ldg_ul(ptr);
-}
-inline __device__ unsigned long long __ldg(const unsigned long long *ptr) {
-  return __nvvm_ldg_ull(ptr);
-}
-inline __device__ float __ldg(const float *ptr) { return __nvvm_ldg_f(ptr); }
-inline __device__ double __ldg(const double *ptr) { return __nvvm_ldg_d(ptr); }
-
-inline __device__ char2 __ldg(const char2 *ptr) {
-  typedef char c2 __attribute__((ext_vector_type(2)));
-  // We can assume that ptr is aligned at least to char2's alignment, but the
-  // load will assume that ptr is aligned to char2's alignment.  This is only
-  // safe if alignof(c2) <= alignof(char2).
-  c2 rv = __nvvm_ldg_c2(reinterpret_cast<const c2 *>(ptr));
-  char2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ char4 __ldg(const char4 *ptr) {
-  typedef char c4 __attribute__((ext_vector_type(4)));
-  c4 rv = __nvvm_ldg_c4(reinterpret_cast<const c4 *>(ptr));
-  char4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ short2 __ldg(const short2 *ptr) {
-  typedef short s2 __attribute__((ext_vector_type(2)));
-  s2 rv = __nvvm_ldg_s2(reinterpret_cast<const s2 *>(ptr));
-  short2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ short4 __ldg(const short4 *ptr) {
-  typedef short s4 __attribute__((ext_vector_type(4)));
-  s4 rv = __nvvm_ldg_s4(reinterpret_cast<const s4 *>(ptr));
-  short4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ int2 __ldg(const int2 *ptr) {
-  typedef int i2 __attribute__((ext_vector_type(2)));
-  i2 rv = __nvvm_ldg_i2(reinterpret_cast<const i2 *>(ptr));
-  int2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ int4 __ldg(const int4 *ptr) {
-  typedef int i4 __attribute__((ext_vector_type(4)));
-  i4 rv = __nvvm_ldg_i4(reinterpret_cast<const i4 *>(ptr));
-  int4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ longlong2 __ldg(const longlong2 *ptr) {
-  typedef long long ll2 __attribute__((ext_vector_type(2)));
-  ll2 rv = __nvvm_ldg_ll2(reinterpret_cast<const ll2 *>(ptr));
-  longlong2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-
-inline __device__ uchar2 __ldg(const uchar2 *ptr) {
-  typedef unsigned char uc2 __attribute__((ext_vector_type(2)));
-  uc2 rv = __nvvm_ldg_uc2(reinterpret_cast<const uc2 *>(ptr));
-  uchar2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ uchar4 __ldg(const uchar4 *ptr) {
-  typedef unsigned char uc4 __attribute__((ext_vector_type(4)));
-  uc4 rv = __nvvm_ldg_uc4(reinterpret_cast<const uc4 *>(ptr));
-  uchar4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ ushort2 __ldg(const ushort2 *ptr) {
-  typedef unsigned short us2 __attribute__((ext_vector_type(2)));
-  us2 rv = __nvvm_ldg_us2(reinterpret_cast<const us2 *>(ptr));
-  ushort2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ ushort4 __ldg(const ushort4 *ptr) {
-  typedef unsigned short us4 __attribute__((ext_vector_type(4)));
-  us4 rv = __nvvm_ldg_us4(reinterpret_cast<const us4 *>(ptr));
-  ushort4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ uint2 __ldg(const uint2 *ptr) {
-  typedef unsigned int ui2 __attribute__((ext_vector_type(2)));
-  ui2 rv = __nvvm_ldg_ui2(reinterpret_cast<const ui2 *>(ptr));
-  uint2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ uint4 __ldg(const uint4 *ptr) {
-  typedef unsigned int ui4 __attribute__((ext_vector_type(4)));
-  ui4 rv = __nvvm_ldg_ui4(reinterpret_cast<const ui4 *>(ptr));
-  uint4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ ulonglong2 __ldg(const ulonglong2 *ptr) {
-  typedef unsigned long long ull2 __attribute__((ext_vector_type(2)));
-  ull2 rv = __nvvm_ldg_ull2(reinterpret_cast<const ull2 *>(ptr));
-  ulonglong2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-
-inline __device__ float2 __ldg(const float2 *ptr) {
-  typedef float f2 __attribute__((ext_vector_type(2)));
-  f2 rv = __nvvm_ldg_f2(reinterpret_cast<const f2 *>(ptr));
-  float2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-inline __device__ float4 __ldg(const float4 *ptr) {
-  typedef float f4 __attribute__((ext_vector_type(4)));
-  f4 rv = __nvvm_ldg_f4(reinterpret_cast<const f4 *>(ptr));
-  float4 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  ret.z = rv[2];
-  ret.w = rv[3];
-  return ret;
-}
-inline __device__ double2 __ldg(const double2 *ptr) {
-  typedef double d2 __attribute__((ext_vector_type(2)));
-  d2 rv = __nvvm_ldg_d2(reinterpret_cast<const d2 *>(ptr));
-  double2 ret;
-  ret.x = rv[0];
-  ret.y = rv[1];
-  return ret;
-}
-
-// TODO: Implement these as intrinsics, so the backend can work its magic on
-// these.  Alternatively, we could implement these as plain C and try to get
-// llvm to recognize the relevant patterns.
-inline __device__ unsigned __funnelshift_l(unsigned low32, unsigned high32,
-                                           unsigned shiftWidth) {
-  unsigned result;
-  asm("shf.l.wrap.b32 %0, %1, %2, %3;"
-      : "=r"(result)
-      : "r"(low32), "r"(high32), "r"(shiftWidth));
-  return result;
-}
-inline __device__ unsigned __funnelshift_lc(unsigned low32, unsigned high32,
-                                            unsigned shiftWidth) {
-  unsigned result;
-  asm("shf.l.clamp.b32 %0, %1, %2, %3;"
-      : "=r"(result)
-      : "r"(low32), "r"(high32), "r"(shiftWidth));
-  return result;
-}
-inline __device__ unsigned __funnelshift_r(unsigned low32, unsigned high32,
-                                           unsigned shiftWidth) {
-  unsigned result;
-  asm("shf.r.wrap.b32 %0, %1, %2, %3;"
-      : "=r"(result)
-      : "r"(low32), "r"(high32), "r"(shiftWidth));
-  return result;
-}
-inline __device__ unsigned __funnelshift_rc(unsigned low32, unsigned high32,
-                                            unsigned shiftWidth) {
-  unsigned ret;
-  asm("shf.r.clamp.b32 %0, %1, %2, %3;"
-      : "=r"(ret)
-      : "r"(low32), "r"(high32), "r"(shiftWidth));
-  return ret;
-}
-
-#endif // !defined(__CUDA_ARCH__) || __CUDA_ARCH__ >= 320
-
-#endif // defined(__CLANG_CUDA_INTRINSICS_H__)
diff --git a/linux-x86/lib64/clang/9.0.7/include/arm_acle.h b/linux-x86/lib64/clang/9.0.7/include/arm_acle.h
deleted file mode 100644
index 08d65fa..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/arm_acle.h
+++ /dev/null
@@ -1,622 +0,0 @@
-/*===---- arm_acle.h - ARM Non-Neon intrinsics -----------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __ARM_ACLE_H
-#define __ARM_ACLE_H
-
-#ifndef __ARM_ACLE
-#error "ACLE intrinsics support not enabled."
-#endif
-
-#include <stdint.h>
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-/* 8 SYNCHRONIZATION, BARRIER AND HINT INTRINSICS */
-/* 8.3 Memory barriers */
-#if !defined(_MSC_VER)
-#define __dmb(i) __builtin_arm_dmb(i)
-#define __dsb(i) __builtin_arm_dsb(i)
-#define __isb(i) __builtin_arm_isb(i)
-#endif
-
-/* 8.4 Hints */
-
-#if !defined(_MSC_VER)
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __wfi(void) {
-  __builtin_arm_wfi();
-}
-
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __wfe(void) {
-  __builtin_arm_wfe();
-}
-
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __sev(void) {
-  __builtin_arm_sev();
-}
-
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __sevl(void) {
-  __builtin_arm_sevl();
-}
-
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __yield(void) {
-  __builtin_arm_yield();
-}
-#endif
-
-#if __ARM_32BIT_STATE
-#define __dbg(t) __builtin_arm_dbg(t)
-#endif
-
-/* 8.5 Swap */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__swp(uint32_t __x, volatile uint32_t *__p) {
-  uint32_t v;
-  do
-    v = __builtin_arm_ldrex(__p);
-  while (__builtin_arm_strex(__x, __p));
-  return v;
-}
-
-/* 8.6 Memory prefetch intrinsics */
-/* 8.6.1 Data prefetch */
-#define __pld(addr) __pldx(0, 0, 0, addr)
-
-#if __ARM_32BIT_STATE
-#define __pldx(access_kind, cache_level, retention_policy, addr) \
-  __builtin_arm_prefetch(addr, access_kind, 1)
-#else
-#define __pldx(access_kind, cache_level, retention_policy, addr) \
-  __builtin_arm_prefetch(addr, access_kind, cache_level, retention_policy, 1)
-#endif
-
-/* 8.6.2 Instruction prefetch */
-#define __pli(addr) __plix(0, 0, addr)
-
-#if __ARM_32BIT_STATE
-#define __plix(cache_level, retention_policy, addr) \
-  __builtin_arm_prefetch(addr, 0, 0)
-#else
-#define __plix(cache_level, retention_policy, addr) \
-  __builtin_arm_prefetch(addr, 0, cache_level, retention_policy, 0)
-#endif
-
-/* 8.7 NOP */
-static __inline__ void __attribute__((__always_inline__, __nodebug__)) __nop(void) {
-  __builtin_arm_nop();
-}
-
-/* 9 DATA-PROCESSING INTRINSICS */
-/* 9.2 Miscellaneous data-processing intrinsics */
-/* ROR */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__ror(uint32_t __x, uint32_t __y) {
-  __y %= 32;
-  if (__y == 0)
-    return __x;
-  return (__x >> __y) | (__x << (32 - __y));
-}
-
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
-__rorll(uint64_t __x, uint32_t __y) {
-  __y %= 64;
-  if (__y == 0)
-    return __x;
-  return (__x >> __y) | (__x << (64 - __y));
-}
-
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
-__rorl(unsigned long __x, uint32_t __y) {
-#if __SIZEOF_LONG__ == 4
-  return __ror(__x, __y);
-#else
-  return __rorll(__x, __y);
-#endif
-}
-
-
-/* CLZ */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__clz(uint32_t __t) {
-  return __builtin_clz(__t);
-}
-
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
-__clzl(unsigned long __t) {
-  return __builtin_clzl(__t);
-}
-
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
-__clzll(uint64_t __t) {
-  return __builtin_clzll(__t);
-}
-
-/* REV */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__rev(uint32_t __t) {
-  return __builtin_bswap32(__t);
-}
-
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
-__revl(unsigned long __t) {
-#if __SIZEOF_LONG__ == 4
-  return __builtin_bswap32(__t);
-#else
-  return __builtin_bswap64(__t);
-#endif
-}
-
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
-__revll(uint64_t __t) {
-  return __builtin_bswap64(__t);
-}
-
-/* REV16 */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__rev16(uint32_t __t) {
-  return __ror(__rev(__t), 16);
-}
-
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
-__rev16ll(uint64_t __t) {
-  return (((uint64_t)__rev16(__t >> 32)) << 32) | __rev16(__t);
-}
-
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
-__rev16l(unsigned long __t) {
-#if __SIZEOF_LONG__ == 4
-    return __rev16(__t);
-#else
-    return __rev16ll(__t);
-#endif
-}
-
-/* REVSH */
-static __inline__ int16_t __attribute__((__always_inline__, __nodebug__))
-__revsh(int16_t __t) {
-  return __builtin_bswap16(__t);
-}
-
-/* RBIT */
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__rbit(uint32_t __t) {
-  return __builtin_arm_rbit(__t);
-}
-
-static __inline__ uint64_t __attribute__((__always_inline__, __nodebug__))
-__rbitll(uint64_t __t) {
-#if __ARM_32BIT_STATE
-  return (((uint64_t)__builtin_arm_rbit(__t)) << 32) |
-         __builtin_arm_rbit(__t >> 32);
-#else
-  return __builtin_arm_rbit64(__t);
-#endif
-}
-
-static __inline__ unsigned long __attribute__((__always_inline__, __nodebug__))
-__rbitl(unsigned long __t) {
-#if __SIZEOF_LONG__ == 4
-  return __rbit(__t);
-#else
-  return __rbitll(__t);
-#endif
-}
-
-/*
- * 9.3 16-bit multiplications
- */
-#if __ARM_FEATURE_DSP
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smulbb(int32_t __a, int32_t __b) {
-  return __builtin_arm_smulbb(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smulbt(int32_t __a, int32_t __b) {
-  return __builtin_arm_smulbt(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smultb(int32_t __a, int32_t __b) {
-  return __builtin_arm_smultb(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smultt(int32_t __a, int32_t __b) {
-  return __builtin_arm_smultt(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smulwb(int32_t __a, int32_t __b) {
-  return __builtin_arm_smulwb(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__,__nodebug__))
-__smulwt(int32_t __a, int32_t __b) {
-  return __builtin_arm_smulwt(__a, __b);
-}
-#endif
-
-/*
- * 9.4 Saturating intrinsics
- *
- * FIXME: Change guard to their corrosponding __ARM_FEATURE flag when Q flag
- * intrinsics are implemented and the flag is enabled.
- */
-/* 9.4.1 Width-specified saturation intrinsics */
-#if __ARM_FEATURE_SAT
-#define __ssat(x, y) __builtin_arm_ssat(x, y)
-#define __usat(x, y) __builtin_arm_usat(x, y)
-#endif
-
-/* 9.4.2 Saturating addition and subtraction intrinsics */
-#if __ARM_FEATURE_DSP
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__qadd(int32_t __t, int32_t __v) {
-  return __builtin_arm_qadd(__t, __v);
-}
-
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__qsub(int32_t __t, int32_t __v) {
-  return __builtin_arm_qsub(__t, __v);
-}
-
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__qdbl(int32_t __t) {
-  return __builtin_arm_qadd(__t, __t);
-}
-#endif
-
-/* 9.4.3 Accumultating multiplications */
-#if __ARM_FEATURE_DSP
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlabb(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlabb(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlabt(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlabt(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlatb(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlatb(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlatt(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlatt(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlawb(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlawb(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlawt(int32_t __a, int32_t __b, int32_t __c) {
-  return __builtin_arm_smlawt(__a, __b, __c);
-}
-#endif
-
-
-/* 9.5.4 Parallel 16-bit saturation */
-#if __ARM_FEATURE_SIMD32
-#define __ssat16(x, y) __builtin_arm_ssat16(x, y)
-#define __usat16(x, y) __builtin_arm_usat16(x, y)
-#endif
-
-/* 9.5.5 Packing and unpacking */
-#if __ARM_FEATURE_SIMD32
-typedef int32_t int8x4_t;
-typedef int32_t int16x2_t;
-typedef uint32_t uint8x4_t;
-typedef uint32_t uint16x2_t;
-
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__sxtab16(int16x2_t __a, int8x4_t __b) {
-  return __builtin_arm_sxtab16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__sxtb16(int8x4_t __a) {
-  return __builtin_arm_sxtb16(__a);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__uxtab16(int16x2_t __a, int8x4_t __b) {
-  return __builtin_arm_uxtab16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__uxtb16(int8x4_t __a) {
-  return __builtin_arm_uxtb16(__a);
-}
-#endif
-
-/* 9.5.6 Parallel selection */
-#if __ARM_FEATURE_SIMD32
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__sel(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_sel(__a, __b);
-}
-#endif
-
-/* 9.5.7 Parallel 8-bit addition and subtraction */
-#if __ARM_FEATURE_SIMD32
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__qadd8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_qadd8(__a, __b);
-}
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__qsub8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_qsub8(__a, __b);
-}
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__sadd8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_sadd8(__a, __b);
-}
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__shadd8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_shadd8(__a, __b);
-}
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__shsub8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_shsub8(__a, __b);
-}
-static __inline__ int8x4_t __attribute__((__always_inline__, __nodebug__))
-__ssub8(int8x4_t __a, int8x4_t __b) {
-  return __builtin_arm_ssub8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__uadd8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_uadd8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__uhadd8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_uhadd8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__uhsub8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_uhsub8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__uqadd8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_uqadd8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__uqsub8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_uqsub8(__a, __b);
-}
-static __inline__ uint8x4_t __attribute__((__always_inline__, __nodebug__))
-__usub8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_usub8(__a, __b);
-}
-#endif
-
-/* 9.5.8 Sum of 8-bit absolute differences */
-#if __ARM_FEATURE_SIMD32
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__usad8(uint8x4_t __a, uint8x4_t __b) {
-  return __builtin_arm_usad8(__a, __b);
-}
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__usada8(uint8x4_t __a, uint8x4_t __b, uint32_t __c) {
-  return __builtin_arm_usada8(__a, __b, __c);
-}
-#endif
-
-/* 9.5.9 Parallel 16-bit addition and subtraction */
-#if __ARM_FEATURE_SIMD32
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__qadd16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_qadd16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__qasx(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_qasx(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__qsax(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_qsax(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__qsub16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_qsub16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__sadd16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_sadd16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__sasx(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_sasx(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__shadd16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_shadd16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__shasx(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_shasx(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__shsax(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_shsax(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__shsub16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_shsub16(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__ssax(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_ssax(__a, __b);
-}
-static __inline__ int16x2_t __attribute__((__always_inline__, __nodebug__))
-__ssub16(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_ssub16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uadd16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uadd16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uasx(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uasx(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uhadd16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uhadd16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uhasx(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uhasx(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uhsax(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uhsax(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uhsub16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uhsub16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uqadd16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uqadd16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uqasx(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uqasx(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uqsax(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uqsax(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__uqsub16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_uqsub16(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__usax(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_usax(__a, __b);
-}
-static __inline__ uint16x2_t __attribute__((__always_inline__, __nodebug__))
-__usub16(uint16x2_t __a, uint16x2_t __b) {
-  return __builtin_arm_usub16(__a, __b);
-}
-#endif
-
-/* 9.5.10 Parallel 16-bit multiplications */
-#if __ARM_FEATURE_SIMD32
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlad(int16x2_t __a, int16x2_t __b, int32_t __c) {
-  return __builtin_arm_smlad(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smladx(int16x2_t __a, int16x2_t __b, int32_t __c) {
-  return __builtin_arm_smladx(__a, __b, __c);
-}
-static __inline__ int64_t __attribute__((__always_inline__, __nodebug__))
-__smlald(int16x2_t __a, int16x2_t __b, int64_t __c) {
-  return __builtin_arm_smlald(__a, __b, __c);
-}
-static __inline__ int64_t __attribute__((__always_inline__, __nodebug__))
-__smlaldx(int16x2_t __a, int16x2_t __b, int64_t __c) {
-  return __builtin_arm_smlaldx(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlsd(int16x2_t __a, int16x2_t __b, int32_t __c) {
-  return __builtin_arm_smlsd(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smlsdx(int16x2_t __a, int16x2_t __b, int32_t __c) {
-  return __builtin_arm_smlsdx(__a, __b, __c);
-}
-static __inline__ int64_t __attribute__((__always_inline__, __nodebug__))
-__smlsld(int16x2_t __a, int16x2_t __b, int64_t __c) {
-  return __builtin_arm_smlsld(__a, __b, __c);
-}
-static __inline__ int64_t __attribute__((__always_inline__, __nodebug__))
-__smlsldx(int16x2_t __a, int16x2_t __b, int64_t __c) {
-  return __builtin_arm_smlsldx(__a, __b, __c);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smuad(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_smuad(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smuadx(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_smuadx(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smusd(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_smusd(__a, __b);
-}
-static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
-__smusdx(int16x2_t __a, int16x2_t __b) {
-  return __builtin_arm_smusdx(__a, __b);
-}
-#endif
-
-/* 9.7 CRC32 intrinsics */
-#if __ARM_FEATURE_CRC32
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32b(uint32_t __a, uint8_t __b) {
-  return __builtin_arm_crc32b(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32h(uint32_t __a, uint16_t __b) {
-  return __builtin_arm_crc32h(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32w(uint32_t __a, uint32_t __b) {
-  return __builtin_arm_crc32w(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32d(uint32_t __a, uint64_t __b) {
-  return __builtin_arm_crc32d(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32cb(uint32_t __a, uint8_t __b) {
-  return __builtin_arm_crc32cb(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32ch(uint32_t __a, uint16_t __b) {
-  return __builtin_arm_crc32ch(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32cw(uint32_t __a, uint32_t __b) {
-  return __builtin_arm_crc32cw(__a, __b);
-}
-
-static __inline__ uint32_t __attribute__((__always_inline__, __nodebug__))
-__crc32cd(uint32_t __a, uint64_t __b) {
-  return __builtin_arm_crc32cd(__a, __b);
-}
-#endif
-
-/* 10.1 Special register intrinsics */
-#define __arm_rsr(sysreg) __builtin_arm_rsr(sysreg)
-#define __arm_rsr64(sysreg) __builtin_arm_rsr64(sysreg)
-#define __arm_rsrp(sysreg) __builtin_arm_rsrp(sysreg)
-#define __arm_wsr(sysreg, v) __builtin_arm_wsr(sysreg, v)
-#define __arm_wsr64(sysreg, v) __builtin_arm_wsr64(sysreg, v)
-#define __arm_wsrp(sysreg, v) __builtin_arm_wsrp(sysreg, v)
-
-// Memory Tagging Extensions (MTE) Intrinsics
-#if __ARM_FEATURE_MEMORY_TAGGING
-#define __arm_mte_create_random_tag(__ptr, __mask)  __builtin_arm_irg(__ptr, __mask)
-#define __arm_mte_increment_tag(__ptr, __tag_offset)  __builtin_arm_addg(__ptr, __tag_offset)
-#define __arm_mte_exclude_tag(__ptr, __excluded)  __builtin_arm_gmi(__ptr, __excluded)
-#define __arm_mte_get_tag(__ptr) __builtin_arm_ldg(__ptr)
-#define __arm_mte_set_tag(__ptr) __builtin_arm_stg(__ptr)
-#define __arm_mte_ptrdiff(__ptra, __ptrb) __builtin_arm_subp(__ptra, __ptrb)
-#endif
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif /* __ARM_ACLE_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/arm_fp16.h b/linux-x86/lib64/clang/9.0.7/include/arm_fp16.h
deleted file mode 100644
index de54465..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/arm_fp16.h
+++ /dev/null
@@ -1,1499 +0,0 @@
-/*===---- arm_fp16.h - ARM FP16 intrinsics ---------------------------------===
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
- * THE SOFTWARE.
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __ARM_FP16_H
-#define __ARM_FP16_H
-
-#include <stdint.h>
-
-typedef __fp16 float16_t;
-#define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
-
-#if defined(__ARM_FEATURE_FP16_SCALAR_ARITHMETIC) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vabdh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vabdh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabdh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vabsh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vabsh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vabsh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vaddh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vaddh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vaddh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcageh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcageh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcageh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcagth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcagth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcagth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcaleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcaleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcaleh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcalth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcalth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcalth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vceqh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vceqh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vceqzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vceqzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vceqzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgeh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcgeh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgeh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcgezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgezh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcgth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcgtzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcgtzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcgtzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vcleh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcleh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vclezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vclezh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclezh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vclth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vclth_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vclth_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcltzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcltzh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcltzh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_n_s16_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_n_s32_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_s64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_n_s64_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_n_u16_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u32_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_n_u32_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_u64_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_n_u64_f16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvth_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvth_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvth_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvth_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvth_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvth_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvth_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtah_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtah_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtah_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtah_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtah_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtah_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtah_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtah_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u32(uint32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u32(uint32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u64(uint64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u64(uint64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_u16(uint16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_u16(uint16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_u16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s32(int32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s32(int32_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s64(int64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s64(int64_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16_t vcvth_f16_s16(int16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
-  return __ret;
-}
-#else
-__ai float16_t vcvth_f16_s16(int16_t __p0) {
-  float16_t __ret;
-  __ret = (float16_t) __builtin_neon_vcvth_f16_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvth_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vcvth_n_f16_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtmh_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtmh_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtmh_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtmh_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtmh_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtmh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtmh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtmh_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtnh_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtnh_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtnh_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtnh_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtnh_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtnh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtnh_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtnh_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vcvtph_s16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvtph_s32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_s64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtph_s64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u16_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vcvtph_u16_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u32_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvtph_u32_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtph_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
-  __ret; \
-})
-#else
-#define vcvtph_u64_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtph_u64_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdivh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vdivh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vdivh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#else
-#define vfmah_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#else
-#define vfmsh_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmsh_f16(__s0, __s1, __s2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmaxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmaxnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vminh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vminnmh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmulh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vmulxh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vnegh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vnegh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vnegh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpeh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrecpeh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpeh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vrecpsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpsh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrecpxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrecpxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrecpxh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndah_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndah_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndah_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndih_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndih_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndih_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndmh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndmh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndmh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndnh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndnh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndnh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndph_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndph_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndph_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrndxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrndxh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrndxh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsqrteh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
-  __ret; \
-})
-#else
-#define vrsqrteh_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrteh_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vrsqrtsh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vrsqrtsh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsqrth_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
-  __ret; \
-})
-#else
-#define vsqrth_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsqrth_f16(__s0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsubh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
-  __ret; \
-})
-#else
-#define vsubh_f16(__p0, __p1) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vsubh_f16(__s0, __s1); \
-  __ret; \
-})
-#endif
-
-#endif
-
-#undef __ai
-
-#endif /* __ARM_FP16_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/arm_neon.h b/linux-x86/lib64/clang/9.0.7/include/arm_neon.h
deleted file mode 100644
index 694bdfc..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/arm_neon.h
+++ /dev/null
@@ -1,73681 +0,0 @@
-/*===---- arm_neon.h - ARM Neon intrinsics ---------------------------------===
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
- * THE SOFTWARE.
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __ARM_NEON_H
-#define __ARM_NEON_H
-
-#if !defined(__ARM_NEON)
-#error "NEON support not enabled"
-#endif
-
-#include <stdint.h>
-
-typedef float float32_t;
-typedef __fp16 float16_t;
-#ifdef __aarch64__
-typedef double float64_t;
-#endif
-
-#ifdef __aarch64__
-typedef uint8_t poly8_t;
-typedef uint16_t poly16_t;
-typedef uint64_t poly64_t;
-typedef __uint128_t poly128_t;
-#else
-typedef int8_t poly8_t;
-typedef int16_t poly16_t;
-#endif
-typedef __attribute__((neon_vector_type(8))) int8_t int8x8_t;
-typedef __attribute__((neon_vector_type(16))) int8_t int8x16_t;
-typedef __attribute__((neon_vector_type(4))) int16_t int16x4_t;
-typedef __attribute__((neon_vector_type(8))) int16_t int16x8_t;
-typedef __attribute__((neon_vector_type(2))) int32_t int32x2_t;
-typedef __attribute__((neon_vector_type(4))) int32_t int32x4_t;
-typedef __attribute__((neon_vector_type(1))) int64_t int64x1_t;
-typedef __attribute__((neon_vector_type(2))) int64_t int64x2_t;
-typedef __attribute__((neon_vector_type(8))) uint8_t uint8x8_t;
-typedef __attribute__((neon_vector_type(16))) uint8_t uint8x16_t;
-typedef __attribute__((neon_vector_type(4))) uint16_t uint16x4_t;
-typedef __attribute__((neon_vector_type(8))) uint16_t uint16x8_t;
-typedef __attribute__((neon_vector_type(2))) uint32_t uint32x2_t;
-typedef __attribute__((neon_vector_type(4))) uint32_t uint32x4_t;
-typedef __attribute__((neon_vector_type(1))) uint64_t uint64x1_t;
-typedef __attribute__((neon_vector_type(2))) uint64_t uint64x2_t;
-typedef __attribute__((neon_vector_type(4))) float16_t float16x4_t;
-typedef __attribute__((neon_vector_type(8))) float16_t float16x8_t;
-typedef __attribute__((neon_vector_type(2))) float32_t float32x2_t;
-typedef __attribute__((neon_vector_type(4))) float32_t float32x4_t;
-#ifdef __aarch64__
-typedef __attribute__((neon_vector_type(1))) float64_t float64x1_t;
-typedef __attribute__((neon_vector_type(2))) float64_t float64x2_t;
-#endif
-typedef __attribute__((neon_polyvector_type(8))) poly8_t poly8x8_t;
-typedef __attribute__((neon_polyvector_type(16))) poly8_t poly8x16_t;
-typedef __attribute__((neon_polyvector_type(4))) poly16_t poly16x4_t;
-typedef __attribute__((neon_polyvector_type(8))) poly16_t poly16x8_t;
-#ifdef __aarch64__
-typedef __attribute__((neon_polyvector_type(1))) poly64_t poly64x1_t;
-typedef __attribute__((neon_polyvector_type(2))) poly64_t poly64x2_t;
-#endif
-
-typedef struct int8x8x2_t {
-  int8x8_t val[2];
-} int8x8x2_t;
-
-typedef struct int8x16x2_t {
-  int8x16_t val[2];
-} int8x16x2_t;
-
-typedef struct int16x4x2_t {
-  int16x4_t val[2];
-} int16x4x2_t;
-
-typedef struct int16x8x2_t {
-  int16x8_t val[2];
-} int16x8x2_t;
-
-typedef struct int32x2x2_t {
-  int32x2_t val[2];
-} int32x2x2_t;
-
-typedef struct int32x4x2_t {
-  int32x4_t val[2];
-} int32x4x2_t;
-
-typedef struct int64x1x2_t {
-  int64x1_t val[2];
-} int64x1x2_t;
-
-typedef struct int64x2x2_t {
-  int64x2_t val[2];
-} int64x2x2_t;
-
-typedef struct uint8x8x2_t {
-  uint8x8_t val[2];
-} uint8x8x2_t;
-
-typedef struct uint8x16x2_t {
-  uint8x16_t val[2];
-} uint8x16x2_t;
-
-typedef struct uint16x4x2_t {
-  uint16x4_t val[2];
-} uint16x4x2_t;
-
-typedef struct uint16x8x2_t {
-  uint16x8_t val[2];
-} uint16x8x2_t;
-
-typedef struct uint32x2x2_t {
-  uint32x2_t val[2];
-} uint32x2x2_t;
-
-typedef struct uint32x4x2_t {
-  uint32x4_t val[2];
-} uint32x4x2_t;
-
-typedef struct uint64x1x2_t {
-  uint64x1_t val[2];
-} uint64x1x2_t;
-
-typedef struct uint64x2x2_t {
-  uint64x2_t val[2];
-} uint64x2x2_t;
-
-typedef struct float16x4x2_t {
-  float16x4_t val[2];
-} float16x4x2_t;
-
-typedef struct float16x8x2_t {
-  float16x8_t val[2];
-} float16x8x2_t;
-
-typedef struct float32x2x2_t {
-  float32x2_t val[2];
-} float32x2x2_t;
-
-typedef struct float32x4x2_t {
-  float32x4_t val[2];
-} float32x4x2_t;
-
-#ifdef __aarch64__
-typedef struct float64x1x2_t {
-  float64x1_t val[2];
-} float64x1x2_t;
-
-typedef struct float64x2x2_t {
-  float64x2_t val[2];
-} float64x2x2_t;
-
-#endif
-typedef struct poly8x8x2_t {
-  poly8x8_t val[2];
-} poly8x8x2_t;
-
-typedef struct poly8x16x2_t {
-  poly8x16_t val[2];
-} poly8x16x2_t;
-
-typedef struct poly16x4x2_t {
-  poly16x4_t val[2];
-} poly16x4x2_t;
-
-typedef struct poly16x8x2_t {
-  poly16x8_t val[2];
-} poly16x8x2_t;
-
-#ifdef __aarch64__
-typedef struct poly64x1x2_t {
-  poly64x1_t val[2];
-} poly64x1x2_t;
-
-typedef struct poly64x2x2_t {
-  poly64x2_t val[2];
-} poly64x2x2_t;
-
-#endif
-typedef struct int8x8x3_t {
-  int8x8_t val[3];
-} int8x8x3_t;
-
-typedef struct int8x16x3_t {
-  int8x16_t val[3];
-} int8x16x3_t;
-
-typedef struct int16x4x3_t {
-  int16x4_t val[3];
-} int16x4x3_t;
-
-typedef struct int16x8x3_t {
-  int16x8_t val[3];
-} int16x8x3_t;
-
-typedef struct int32x2x3_t {
-  int32x2_t val[3];
-} int32x2x3_t;
-
-typedef struct int32x4x3_t {
-  int32x4_t val[3];
-} int32x4x3_t;
-
-typedef struct int64x1x3_t {
-  int64x1_t val[3];
-} int64x1x3_t;
-
-typedef struct int64x2x3_t {
-  int64x2_t val[3];
-} int64x2x3_t;
-
-typedef struct uint8x8x3_t {
-  uint8x8_t val[3];
-} uint8x8x3_t;
-
-typedef struct uint8x16x3_t {
-  uint8x16_t val[3];
-} uint8x16x3_t;
-
-typedef struct uint16x4x3_t {
-  uint16x4_t val[3];
-} uint16x4x3_t;
-
-typedef struct uint16x8x3_t {
-  uint16x8_t val[3];
-} uint16x8x3_t;
-
-typedef struct uint32x2x3_t {
-  uint32x2_t val[3];
-} uint32x2x3_t;
-
-typedef struct uint32x4x3_t {
-  uint32x4_t val[3];
-} uint32x4x3_t;
-
-typedef struct uint64x1x3_t {
-  uint64x1_t val[3];
-} uint64x1x3_t;
-
-typedef struct uint64x2x3_t {
-  uint64x2_t val[3];
-} uint64x2x3_t;
-
-typedef struct float16x4x3_t {
-  float16x4_t val[3];
-} float16x4x3_t;
-
-typedef struct float16x8x3_t {
-  float16x8_t val[3];
-} float16x8x3_t;
-
-typedef struct float32x2x3_t {
-  float32x2_t val[3];
-} float32x2x3_t;
-
-typedef struct float32x4x3_t {
-  float32x4_t val[3];
-} float32x4x3_t;
-
-#ifdef __aarch64__
-typedef struct float64x1x3_t {
-  float64x1_t val[3];
-} float64x1x3_t;
-
-typedef struct float64x2x3_t {
-  float64x2_t val[3];
-} float64x2x3_t;
-
-#endif
-typedef struct poly8x8x3_t {
-  poly8x8_t val[3];
-} poly8x8x3_t;
-
-typedef struct poly8x16x3_t {
-  poly8x16_t val[3];
-} poly8x16x3_t;
-
-typedef struct poly16x4x3_t {
-  poly16x4_t val[3];
-} poly16x4x3_t;
-
-typedef struct poly16x8x3_t {
-  poly16x8_t val[3];
-} poly16x8x3_t;
-
-#ifdef __aarch64__
-typedef struct poly64x1x3_t {
-  poly64x1_t val[3];
-} poly64x1x3_t;
-
-typedef struct poly64x2x3_t {
-  poly64x2_t val[3];
-} poly64x2x3_t;
-
-#endif
-typedef struct int8x8x4_t {
-  int8x8_t val[4];
-} int8x8x4_t;
-
-typedef struct int8x16x4_t {
-  int8x16_t val[4];
-} int8x16x4_t;
-
-typedef struct int16x4x4_t {
-  int16x4_t val[4];
-} int16x4x4_t;
-
-typedef struct int16x8x4_t {
-  int16x8_t val[4];
-} int16x8x4_t;
-
-typedef struct int32x2x4_t {
-  int32x2_t val[4];
-} int32x2x4_t;
-
-typedef struct int32x4x4_t {
-  int32x4_t val[4];
-} int32x4x4_t;
-
-typedef struct int64x1x4_t {
-  int64x1_t val[4];
-} int64x1x4_t;
-
-typedef struct int64x2x4_t {
-  int64x2_t val[4];
-} int64x2x4_t;
-
-typedef struct uint8x8x4_t {
-  uint8x8_t val[4];
-} uint8x8x4_t;
-
-typedef struct uint8x16x4_t {
-  uint8x16_t val[4];
-} uint8x16x4_t;
-
-typedef struct uint16x4x4_t {
-  uint16x4_t val[4];
-} uint16x4x4_t;
-
-typedef struct uint16x8x4_t {
-  uint16x8_t val[4];
-} uint16x8x4_t;
-
-typedef struct uint32x2x4_t {
-  uint32x2_t val[4];
-} uint32x2x4_t;
-
-typedef struct uint32x4x4_t {
-  uint32x4_t val[4];
-} uint32x4x4_t;
-
-typedef struct uint64x1x4_t {
-  uint64x1_t val[4];
-} uint64x1x4_t;
-
-typedef struct uint64x2x4_t {
-  uint64x2_t val[4];
-} uint64x2x4_t;
-
-typedef struct float16x4x4_t {
-  float16x4_t val[4];
-} float16x4x4_t;
-
-typedef struct float16x8x4_t {
-  float16x8_t val[4];
-} float16x8x4_t;
-
-typedef struct float32x2x4_t {
-  float32x2_t val[4];
-} float32x2x4_t;
-
-typedef struct float32x4x4_t {
-  float32x4_t val[4];
-} float32x4x4_t;
-
-#ifdef __aarch64__
-typedef struct float64x1x4_t {
-  float64x1_t val[4];
-} float64x1x4_t;
-
-typedef struct float64x2x4_t {
-  float64x2_t val[4];
-} float64x2x4_t;
-
-#endif
-typedef struct poly8x8x4_t {
-  poly8x8_t val[4];
-} poly8x8x4_t;
-
-typedef struct poly8x16x4_t {
-  poly8x16_t val[4];
-} poly8x16x4_t;
-
-typedef struct poly16x4x4_t {
-  poly16x4_t val[4];
-} poly16x4x4_t;
-
-typedef struct poly16x8x4_t {
-  poly16x8_t val[4];
-} poly16x8x4_t;
-
-#ifdef __aarch64__
-typedef struct poly64x1x4_t {
-  poly64x1_t val[4];
-} poly64x1x4_t;
-
-typedef struct poly64x2x4_t {
-  poly64x2_t val[4];
-} poly64x2x4_t;
-
-#endif
-
-#define __ai static __inline__ __attribute__((__always_inline__, __nodebug__))
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x16_t __noswap_vabdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vabdq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vabdq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x16_t __noswap_vabdq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vabdq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vabdq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vabdq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vabd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vabd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vabd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vabd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vabd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vabd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vabd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vabd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vabd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vabd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vabsq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vabsq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vabsq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vabsq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabsq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vabsq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabsq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vabsq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vabs_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vabs_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vabs_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vabs_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vabs_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vabs_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vabs_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vabs_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float32x4_t vaddq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float32x2_t vadd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vaddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vaddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vaddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vaddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vaddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vaddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vaddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vandq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vandq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vandq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vandq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vandq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vandq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vandq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vandq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vand_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vand_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vand_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vand_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vand_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vand_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vand_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 & __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vand_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 & __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vbicq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vbicq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vbicq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vbicq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vbicq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vbicq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vbicq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vbicq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vbic_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vbic_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vbic_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vbic_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vbic_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vbic_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vbic_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 & ~__p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vbic_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 & ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vbsl_p8(uint8x8_t __p0, poly8x8_t __p1, poly8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 5);
-  return __ret;
-}
-#else
-__ai poly16x4_t vbsl_p16(uint16x4_t __p0, poly16x4_t __p1, poly16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vbslq_p8(uint8x16_t __p0, poly8x16_t __p1, poly8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 37);
-  return __ret;
-}
-#else
-__ai poly16x8_t vbslq_p16(uint16x8_t __p0, poly16x8_t __p1, poly16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 37);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vbslq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vbslq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vbslq_u64(uint64x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vbslq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vbslq_s8(uint8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vbslq_f32(uint32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vbslq_s32(uint32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vbslq_s64(uint64x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vbslq_s16(uint16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vbsl_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vbsl_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vbsl_u64(uint64x1_t __p0, uint64x1_t __p1, uint64x1_t __p2) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vbsl_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vbsl_s8(uint8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vbsl_f32(uint32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vbsl_s32(uint32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vbsl_s64(uint64x1_t __p0, int64x1_t __p1, int64x1_t __p2) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vbsl_s16(uint16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcageq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcage_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcage_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcagtq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcagt_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcagt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcaleq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcale_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcale_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcaltq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcalt_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcalt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceq_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceq_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceq_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceq_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceq_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceq_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceq_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceq_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgeq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgeq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgeq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgeq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgeq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgeq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgeq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcge_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcge_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcge_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcge_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcge_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcge_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcge_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgtq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgtq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgtq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgtq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgtq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgtq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgtq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcgt_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgt_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgt_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcgt_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgt_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgt_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgt_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcleq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcleq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcleq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcleq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcleq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcleq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcleq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcle_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcle_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcle_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcle_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcle_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcle_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcle_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vclsq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vclsq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vclsq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vclsq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vclsq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vclsq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vclsq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vcls_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vcls_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcls_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcls_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcls_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcls_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcls_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcltq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcltq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcltq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcltq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcltq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcltq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcltq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vclt_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclt_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclt_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint8x8_t vclt_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclt_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclt_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclt_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vclzq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vclzq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vclzq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vclzq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vclzq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vclzq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vclzq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vclzq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vclzq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vclzq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vclz_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclz_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclz_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vclz_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vclz_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vclz_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vclz_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vclz_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vclz_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vclz_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vcnt_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vcntq_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcntq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vcntq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vcntq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vcntq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcnt_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vcnt_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vcnt_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vcnt_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai poly8x16_t vcombine_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai poly16x8_t vcombine_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x16_t __noswap_vcombine_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vcombine_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcombine_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vcombine_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai int8x16_t vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x16_t __noswap_vcombine_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai float32x4_t vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vcombine_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai float16x8_t vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x8_t __noswap_vcombine_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai int32x4_t vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vcombine_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  return __ret;
-}
-#else
-__ai int64x2_t vcombine_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai int16x8_t vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vcombine_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vcreate_p8(uint64_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vcreate_p8(uint64_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vcreate_p16(uint64_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vcreate_p16(uint64_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcreate_u8(uint64_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcreate_u8(uint64_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcreate_u32(uint64_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcreate_u32(uint64_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcreate_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcreate_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcreate_u16(uint64_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcreate_u16(uint64_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vcreate_s8(uint64_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vcreate_s8(uint64_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vcreate_f32(uint64_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vcreate_f32(uint64_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vcreate_f16(uint64_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vcreate_f16(uint64_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcreate_s32(uint64_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vcreate_s32(uint64_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcreate_s64(uint64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vcreate_s64(uint64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcreate_s16(uint64_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vcreate_s16(uint64_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai float32x4_t vcvtq_f32_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai float32x4_t vcvtq_f32_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvtq_f32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai float32x2_t vcvt_f32_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai float32x2_t vcvt_f32_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vcvtq_n_f32_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vcvt_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vcvt_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vcvt_n_f32_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vcvtq_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vcvtq_n_s32_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vcvt_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vcvt_n_s32_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vcvtq_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vcvtq_n_u32_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vcvt_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vcvt_n_u32_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vcvtq_s32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtq_s32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcvt_s32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvt_s32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcvtq_u32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtq_u32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcvt_u32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvt_u32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly8x8_t vdup_n_p8(poly8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly16x4_t vdup_n_p16(poly16_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly8x16_t vdupq_n_p8(poly8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly16x8_t vdupq_n_p16(poly16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint8x16_t vdupq_n_u8(uint8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint32x4_t vdupq_n_u32(uint32_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai uint64x2_t vdupq_n_u64(uint64_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint16x8_t vdupq_n_u16(uint16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vdupq_n_s8(int8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int8x16_t vdupq_n_s8(int8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vdupq_n_f32(float32_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai float32x4_t vdupq_n_f32(float32_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vdupq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vdupq_n_s32(int32_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int32x4_t vdupq_n_s32(int32_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vdupq_n_s64(int64_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai int64x2_t vdupq_n_s64(int64_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vdupq_n_s16(int16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int16x8_t vdupq_n_s16(int16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint8x8_t vdup_n_u8(uint8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai uint32x2_t vdup_n_u32(uint32_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai uint64x1_t vdup_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint16x4_t vdup_n_u16(uint16_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vdup_n_s8(int8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int8x8_t vdup_n_s8(int8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vdup_n_f32(float32_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai float32x2_t vdup_n_f32(float32_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vdup_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vdup_n_s32(int32_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai int32x2_t vdup_n_s32(int32_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vdup_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai int64x1_t vdup_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vdup_n_s16(int16_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int16x4_t vdup_n_s16(int16_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t veorq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t veorq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t veorq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t veorq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t veorq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int64x2_t veorq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t veorq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t veor_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t veor_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t veor_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t veor_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t veor_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t veor_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int64x1_t veor_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 ^ __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t veor_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 ^ __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
-  __ret; \
-})
-#else
-#define vext_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
-  __ret; \
-})
-#else
-#define vext_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
-  __ret; \
-})
-#else
-#define vextq_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
-  __ret; \
-})
-#else
-#define vextq_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vextq_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vextq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vextq_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vextq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vextq_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 41); \
-  __ret; \
-})
-#else
-#define vextq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vextq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vextq_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vextq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vext_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vext_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vext_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vext_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vext_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 9); \
-  __ret; \
-})
-#else
-#define vext_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vext_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vext_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vext_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai poly8x8_t vget_high_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai poly8x8_t __noswap_vget_high_p8(poly8x16_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai poly16x4_t vget_high_p16(poly16x8_t __p0) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai uint8x8_t vget_high_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vget_high_u8(uint8x16_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#else
-__ai uint32x2_t vget_high_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vget_high_u32(uint32x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vget_high_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai uint16x4_t vget_high_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vget_high_u16(uint16x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vget_high_s8(int8x16_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#else
-__ai int8x8_t vget_high_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 8, 9, 10, 11, 12, 13, 14, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vget_high_s8(int8x16_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 8, 9, 10, 11, 12, 13, 14, 15);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vget_high_f32(float32x4_t __p0) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#else
-__ai float32x2_t vget_high_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vget_high_f32(float32x4_t __p0) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vget_high_f16(float16x8_t __p0) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai float16x4_t vget_high_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x4_t __noswap_vget_high_f16(float16x8_t __p0) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vget_high_s32(int32x4_t __p0) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#else
-__ai int32x2_t vget_high_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vget_high_s32(int32x4_t __p0) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 2, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vget_high_s64(int64x2_t __p0) {
-  int64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1);
-  return __ret;
-}
-#else
-__ai int64x1_t vget_high_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vget_high_s16(int16x8_t __p0) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai int16x4_t vget_high_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vget_high_s16(int16x8_t __p0) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 4, 5, 6, 7);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vgetq_lane_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vgetq_lane_f32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vgetq_lane_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vgetq_lane_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vget_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vget_lane_f32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vget_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vget_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai poly8x8_t vget_low_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai poly16x4_t vget_low_p16(poly16x8_t __p0) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai uint8x8_t vget_low_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
-  return __ret;
-}
-#else
-__ai uint32x2_t vget_low_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vget_low_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai uint16x4_t vget_low_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vget_low_s8(int8x16_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3, 4, 5, 6, 7);
-  return __ret;
-}
-#else
-__ai int8x8_t vget_low_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3, 4, 5, 6, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vget_low_f32(float32x4_t __p0) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
-  return __ret;
-}
-#else
-__ai float32x2_t vget_low_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vget_low_f16(float16x8_t __p0) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai float16x4_t vget_low_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vget_low_s32(int32x4_t __p0) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1);
-  return __ret;
-}
-#else
-__ai int32x2_t vget_low_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vget_low_s64(int64x2_t __p0) {
-  int64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0);
-  return __ret;
-}
-#else
-__ai int64x1_t vget_low_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vget_low_s16(int16x8_t __p0) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0, 1, 2, 3);
-  return __ret;
-}
-#else
-__ai int16x4_t vget_low_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0, 1, 2, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vhadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vhadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vhadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vhsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vhsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vhsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vhsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vhsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vhsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vhsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vhsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vhsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vhsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vhsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vhsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vhsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vhsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p8(__p0) __extension__ ({ \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
-  __ret; \
-})
-#else
-#define vld1_p8(__p0) __extension__ ({ \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_v(__p0, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p16(__p0) __extension__ ({ \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
-  __ret; \
-})
-#else
-#define vld1_p16(__p0) __extension__ ({ \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_v(__p0, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p8(__p0) __extension__ ({ \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
-  __ret; \
-})
-#else
-#define vld1q_p8(__p0) __extension__ ({ \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_v(__p0, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p16(__p0) __extension__ ({ \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
-  __ret; \
-})
-#else
-#define vld1q_p16(__p0) __extension__ ({ \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_v(__p0, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u8(__p0) __extension__ ({ \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
-  __ret; \
-})
-#else
-#define vld1q_u8(__p0) __extension__ ({ \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_v(__p0, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u32(__p0) __extension__ ({ \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
-  __ret; \
-})
-#else
-#define vld1q_u32(__p0) __extension__ ({ \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_v(__p0, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u64(__p0) __extension__ ({ \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
-  __ret; \
-})
-#else
-#define vld1q_u64(__p0) __extension__ ({ \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_v(__p0, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u16(__p0) __extension__ ({ \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
-  __ret; \
-})
-#else
-#define vld1q_u16(__p0) __extension__ ({ \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_v(__p0, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s8(__p0) __extension__ ({ \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
-  __ret; \
-})
-#else
-#define vld1q_s8(__p0) __extension__ ({ \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_v(__p0, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f32(__p0) __extension__ ({ \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
-  __ret; \
-})
-#else
-#define vld1q_f32(__p0) __extension__ ({ \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_v(__p0, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s32(__p0) __extension__ ({ \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
-  __ret; \
-})
-#else
-#define vld1q_s32(__p0) __extension__ ({ \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_v(__p0, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s64(__p0) __extension__ ({ \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
-  __ret; \
-})
-#else
-#define vld1q_s64(__p0) __extension__ ({ \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_v(__p0, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s16(__p0) __extension__ ({ \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
-  __ret; \
-})
-#else
-#define vld1q_s16(__p0) __extension__ ({ \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_v(__p0, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u8(__p0) __extension__ ({ \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
-  __ret; \
-})
-#else
-#define vld1_u8(__p0) __extension__ ({ \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_v(__p0, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u32(__p0) __extension__ ({ \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
-  __ret; \
-})
-#else
-#define vld1_u32(__p0) __extension__ ({ \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_v(__p0, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
-  __ret; \
-})
-#else
-#define vld1_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u16(__p0) __extension__ ({ \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
-  __ret; \
-})
-#else
-#define vld1_u16(__p0) __extension__ ({ \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_v(__p0, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s8(__p0) __extension__ ({ \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
-  __ret; \
-})
-#else
-#define vld1_s8(__p0) __extension__ ({ \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_v(__p0, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f32(__p0) __extension__ ({ \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
-  __ret; \
-})
-#else
-#define vld1_f32(__p0) __extension__ ({ \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_v(__p0, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s32(__p0) __extension__ ({ \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
-  __ret; \
-})
-#else
-#define vld1_s32(__p0) __extension__ ({ \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_v(__p0, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
-  __ret; \
-})
-#else
-#define vld1_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s16(__p0) __extension__ ({ \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
-  __ret; \
-})
-#else
-#define vld1_s16(__p0) __extension__ ({ \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_v(__p0, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_p8(__p0) __extension__ ({ \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
-  __ret; \
-})
-#else
-#define vld1_dup_p8(__p0) __extension__ ({ \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_dup_v(__p0, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_p16(__p0) __extension__ ({ \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
-  __ret; \
-})
-#else
-#define vld1_dup_p16(__p0) __extension__ ({ \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_dup_v(__p0, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_p8(__p0) __extension__ ({ \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
-  __ret; \
-})
-#else
-#define vld1q_dup_p8(__p0) __extension__ ({ \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_dup_v(__p0, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_p16(__p0) __extension__ ({ \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
-  __ret; \
-})
-#else
-#define vld1q_dup_p16(__p0) __extension__ ({ \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_dup_v(__p0, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_u8(__p0) __extension__ ({ \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
-  __ret; \
-})
-#else
-#define vld1q_dup_u8(__p0) __extension__ ({ \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_dup_v(__p0, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_u32(__p0) __extension__ ({ \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
-  __ret; \
-})
-#else
-#define vld1q_dup_u32(__p0) __extension__ ({ \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_dup_v(__p0, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_u64(__p0) __extension__ ({ \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
-  __ret; \
-})
-#else
-#define vld1q_dup_u64(__p0) __extension__ ({ \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_dup_v(__p0, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_u16(__p0) __extension__ ({ \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
-  __ret; \
-})
-#else
-#define vld1q_dup_u16(__p0) __extension__ ({ \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_dup_v(__p0, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_s8(__p0) __extension__ ({ \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
-  __ret; \
-})
-#else
-#define vld1q_dup_s8(__p0) __extension__ ({ \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_dup_v(__p0, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_f32(__p0) __extension__ ({ \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
-  __ret; \
-})
-#else
-#define vld1q_dup_f32(__p0) __extension__ ({ \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_dup_v(__p0, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_s32(__p0) __extension__ ({ \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
-  __ret; \
-})
-#else
-#define vld1q_dup_s32(__p0) __extension__ ({ \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_dup_v(__p0, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_s64(__p0) __extension__ ({ \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
-  __ret; \
-})
-#else
-#define vld1q_dup_s64(__p0) __extension__ ({ \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_dup_v(__p0, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_s16(__p0) __extension__ ({ \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
-  __ret; \
-})
-#else
-#define vld1q_dup_s16(__p0) __extension__ ({ \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_dup_v(__p0, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_u8(__p0) __extension__ ({ \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
-  __ret; \
-})
-#else
-#define vld1_dup_u8(__p0) __extension__ ({ \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_dup_v(__p0, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_u32(__p0) __extension__ ({ \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
-  __ret; \
-})
-#else
-#define vld1_dup_u32(__p0) __extension__ ({ \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_dup_v(__p0, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
-  __ret; \
-})
-#else
-#define vld1_dup_u64(__p0) __extension__ ({ \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_dup_v(__p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_u16(__p0) __extension__ ({ \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
-  __ret; \
-})
-#else
-#define vld1_dup_u16(__p0) __extension__ ({ \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_dup_v(__p0, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_s8(__p0) __extension__ ({ \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
-  __ret; \
-})
-#else
-#define vld1_dup_s8(__p0) __extension__ ({ \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_dup_v(__p0, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_f32(__p0) __extension__ ({ \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
-  __ret; \
-})
-#else
-#define vld1_dup_f32(__p0) __extension__ ({ \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_dup_v(__p0, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_s32(__p0) __extension__ ({ \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
-  __ret; \
-})
-#else
-#define vld1_dup_s32(__p0) __extension__ ({ \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_dup_v(__p0, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
-  __ret; \
-})
-#else
-#define vld1_dup_s64(__p0) __extension__ ({ \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_dup_v(__p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_s16(__p0) __extension__ ({ \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
-  __ret; \
-})
-#else
-#define vld1_dup_s16(__p0) __extension__ ({ \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_dup_v(__p0, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
-  __ret; \
-})
-#else
-#define vld1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
-  __ret; \
-})
-#else
-#define vld1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
-  __ret; \
-})
-#else
-#define vld1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
-  __ret; \
-})
-#else
-#define vld1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vld1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vld1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vld1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vld1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vld1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
-  __ret; \
-})
-#else
-#define vld1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vld1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vld1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vld1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vld1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vld1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vld1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vld1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vld1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
-  __ret; \
-})
-#else
-#define vld1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vld1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vld1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vld1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p8_x2(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld1_p8_x2(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p16_x2(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld1_p16_x2(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p8_x2(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld1q_p8_x2(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p16_x2(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld1q_p16_x2(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u8_x2(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld1q_u8_x2(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u32_x2(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld1q_u32_x2(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u64_x2(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld1q_u64_x2(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u16_x2(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld1q_u16_x2(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s8_x2(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld1q_s8_x2(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f32_x2(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld1q_f32_x2(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s32_x2(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld1q_s32_x2(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s64_x2(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld1q_s64_x2(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s16_x2(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld1q_s16_x2(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u8_x2(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld1_u8_x2(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u32_x2(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld1_u32_x2(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u64_x2(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld1_u64_x2(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u16_x2(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld1_u16_x2(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s8_x2(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld1_s8_x2(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f32_x2(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld1_f32_x2(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s32_x2(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld1_s32_x2(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s64_x2(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld1_s64_x2(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s16_x2(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld1_s16_x2(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p8_x3(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld1_p8_x3(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p16_x3(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld1_p16_x3(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p8_x3(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld1q_p8_x3(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p16_x3(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld1q_p16_x3(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u8_x3(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld1q_u8_x3(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u32_x3(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld1q_u32_x3(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u64_x3(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld1q_u64_x3(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u16_x3(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld1q_u16_x3(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s8_x3(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld1q_s8_x3(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f32_x3(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld1q_f32_x3(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s32_x3(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld1q_s32_x3(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s64_x3(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld1q_s64_x3(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s16_x3(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld1q_s16_x3(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u8_x3(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld1_u8_x3(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u32_x3(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld1_u32_x3(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u64_x3(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld1_u64_x3(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u16_x3(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld1_u16_x3(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s8_x3(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld1_s8_x3(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f32_x3(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld1_f32_x3(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s32_x3(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld1_s32_x3(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s64_x3(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld1_s64_x3(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s16_x3(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld1_s16_x3(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p8_x4(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld1_p8_x4(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p16_x4(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld1_p16_x4(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p8_x4(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld1q_p8_x4(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p16_x4(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld1q_p16_x4(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u8_x4(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld1q_u8_x4(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u32_x4(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld1q_u32_x4(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u64_x4(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld1q_u64_x4(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_u16_x4(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld1q_u16_x4(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s8_x4(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld1q_s8_x4(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f32_x4(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld1q_f32_x4(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s32_x4(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld1q_s32_x4(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s64_x4(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld1q_s64_x4(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_s16_x4(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld1q_s16_x4(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u8_x4(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld1_u8_x4(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u32_x4(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld1_u32_x4(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u64_x4(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld1_u64_x4(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_u16_x4(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld1_u16_x4(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s8_x4(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld1_s8_x4(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f32_x4(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld1_f32_x4(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s32_x4(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld1_s32_x4(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s64_x4(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld1_s64_x4(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_s16_x4(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld1_s16_x4(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_p8(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld2_p8(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_p16(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld2_p16(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_p8(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld2q_p8(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_p16(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld2q_p16(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_u8(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld2q_u8(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_u32(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld2q_u32(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_u16(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld2q_u16(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_s8(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld2q_s8(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_f32(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld2q_f32(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_s32(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld2q_s32(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_s16(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld2q_s16(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_u8(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld2_u8(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_u32(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld2_u32(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld2_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_u16(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld2_u16(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_s8(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld2_s8(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_f32(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld2_f32(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_s32(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld2_s32(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld2_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_s16(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld2_s16(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_p8(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld2_dup_p8(__p0) __extension__ ({ \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_p16(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld2_dup_p16(__p0) __extension__ ({ \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld2q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld2q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld2q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld2q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld2q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld2q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_s8(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld2q_dup_s8(__p0) __extension__ ({ \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_f32(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld2q_dup_f32(__p0) __extension__ ({ \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_s32(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld2q_dup_s32(__p0) __extension__ ({ \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_s64(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld2q_dup_s64(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_s16(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld2q_dup_s16(__p0) __extension__ ({ \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_u8(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld2_dup_u8(__p0) __extension__ ({ \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_u32(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld2_dup_u32(__p0) __extension__ ({ \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld2_dup_u64(__p0) __extension__ ({ \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_u16(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld2_dup_u16(__p0) __extension__ ({ \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_s8(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld2_dup_s8(__p0) __extension__ ({ \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_f32(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld2_dup_f32(__p0) __extension__ ({ \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_s32(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld2_dup_s32(__p0) __extension__ ({ \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld2_dup_s64(__p0) __extension__ ({ \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_s16(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld2_dup_s16(__p0) __extension__ ({ \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
-  __ret; \
-})
-#else
-#define vld2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  poly8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
-  __ret; \
-})
-#else
-#define vld2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  poly16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  poly16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
-  __ret; \
-})
-#else
-#define vld2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  poly16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
-  __ret; \
-})
-#else
-#define vld2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  uint32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  uint32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
-  __ret; \
-})
-#else
-#define vld2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  uint16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 41); \
-  __ret; \
-})
-#else
-#define vld2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  float32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  float32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 34); \
-  __ret; \
-})
-#else
-#define vld2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  int32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  int32x4x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 33); \
-  __ret; \
-})
-#else
-#define vld2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  int16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
-  __ret; \
-})
-#else
-#define vld2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  uint8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
-  __ret; \
-})
-#else
-#define vld2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  uint32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  uint32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
-  __ret; \
-})
-#else
-#define vld2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  uint16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  uint16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
-  __ret; \
-})
-#else
-#define vld2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  int8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 9); \
-  __ret; \
-})
-#else
-#define vld2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  float32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  float32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 2); \
-  __ret; \
-})
-#else
-#define vld2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  int32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  int32x2x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 1); \
-  __ret; \
-})
-#else
-#define vld2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  int16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  int16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_p8(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld3_p8(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_p16(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld3_p16(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_p8(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld3q_p8(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_p16(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld3q_p16(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_u8(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld3q_u8(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_u32(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld3q_u32(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_u16(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld3q_u16(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_s8(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld3q_s8(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_f32(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld3q_f32(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_s32(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld3q_s32(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_s16(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld3q_s16(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_u8(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld3_u8(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_u32(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld3_u32(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld3_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_u16(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld3_u16(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_s8(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld3_s8(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_f32(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld3_f32(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_s32(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld3_s32(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld3_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_s16(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld3_s16(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_p8(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld3_dup_p8(__p0) __extension__ ({ \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_p16(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld3_dup_p16(__p0) __extension__ ({ \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld3q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld3q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld3q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld3q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld3q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld3q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_s8(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld3q_dup_s8(__p0) __extension__ ({ \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_f32(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld3q_dup_f32(__p0) __extension__ ({ \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_s32(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld3q_dup_s32(__p0) __extension__ ({ \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_s64(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld3q_dup_s64(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_s16(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld3q_dup_s16(__p0) __extension__ ({ \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_u8(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld3_dup_u8(__p0) __extension__ ({ \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_u32(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld3_dup_u32(__p0) __extension__ ({ \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld3_dup_u64(__p0) __extension__ ({ \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_u16(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld3_dup_u16(__p0) __extension__ ({ \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_s8(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld3_dup_s8(__p0) __extension__ ({ \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_f32(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld3_dup_f32(__p0) __extension__ ({ \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_s32(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld3_dup_s32(__p0) __extension__ ({ \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld3_dup_s64(__p0) __extension__ ({ \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_s16(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld3_dup_s16(__p0) __extension__ ({ \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
-  __ret; \
-})
-#else
-#define vld3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  poly8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
-  __ret; \
-})
-#else
-#define vld3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  poly16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  poly16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
-  __ret; \
-})
-#else
-#define vld3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  poly16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
-  __ret; \
-})
-#else
-#define vld3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  uint32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  uint32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
-  __ret; \
-})
-#else
-#define vld3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  uint16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 41); \
-  __ret; \
-})
-#else
-#define vld3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  float32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  float32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 34); \
-  __ret; \
-})
-#else
-#define vld3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  int32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  int32x4x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 33); \
-  __ret; \
-})
-#else
-#define vld3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  int16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
-  __ret; \
-})
-#else
-#define vld3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  uint8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
-  __ret; \
-})
-#else
-#define vld3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  uint32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  uint32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
-  __ret; \
-})
-#else
-#define vld3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  uint16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  uint16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
-  __ret; \
-})
-#else
-#define vld3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  int8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 9); \
-  __ret; \
-})
-#else
-#define vld3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  float32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  float32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 2); \
-  __ret; \
-})
-#else
-#define vld3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  int32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  int32x2x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 1); \
-  __ret; \
-})
-#else
-#define vld3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  int16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  int16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_p8(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld4_p8(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_p16(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld4_p16(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_p8(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld4q_p8(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_p16(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld4q_p16(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_u8(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld4q_u8(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_u32(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld4q_u32(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_u16(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld4q_u16(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_s8(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld4q_s8(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_f32(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld4q_f32(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_s32(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld4q_s32(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_s16(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld4q_s16(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_u8(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld4_u8(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_u32(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld4_u32(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld4_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_u16(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld4_u16(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_s8(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld4_s8(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_f32(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld4_f32(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_s32(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld4_s32(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld4_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_s16(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld4_s16(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_p8(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
-  __ret; \
-})
-#else
-#define vld4_dup_p8(__p0) __extension__ ({ \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_p16(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
-  __ret; \
-})
-#else
-#define vld4_dup_p16(__p0) __extension__ ({ \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
-  __ret; \
-})
-#else
-#define vld4q_dup_p8(__p0) __extension__ ({ \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
-  __ret; \
-})
-#else
-#define vld4q_dup_p16(__p0) __extension__ ({ \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
-  __ret; \
-})
-#else
-#define vld4q_dup_u8(__p0) __extension__ ({ \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
-  __ret; \
-})
-#else
-#define vld4q_dup_u32(__p0) __extension__ ({ \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld4q_dup_u64(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
-  __ret; \
-})
-#else
-#define vld4q_dup_u16(__p0) __extension__ ({ \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_s8(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
-  __ret; \
-})
-#else
-#define vld4q_dup_s8(__p0) __extension__ ({ \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_f32(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
-  __ret; \
-})
-#else
-#define vld4q_dup_f32(__p0) __extension__ ({ \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_s32(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
-  __ret; \
-})
-#else
-#define vld4q_dup_s32(__p0) __extension__ ({ \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_s64(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld4q_dup_s64(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_s16(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
-  __ret; \
-})
-#else
-#define vld4q_dup_s16(__p0) __extension__ ({ \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_u8(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
-  __ret; \
-})
-#else
-#define vld4_dup_u8(__p0) __extension__ ({ \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_u32(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
-  __ret; \
-})
-#else
-#define vld4_dup_u32(__p0) __extension__ ({ \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#else
-#define vld4_dup_u64(__p0) __extension__ ({ \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_u16(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
-  __ret; \
-})
-#else
-#define vld4_dup_u16(__p0) __extension__ ({ \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_s8(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
-  __ret; \
-})
-#else
-#define vld4_dup_s8(__p0) __extension__ ({ \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_f32(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
-  __ret; \
-})
-#else
-#define vld4_dup_f32(__p0) __extension__ ({ \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_s32(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
-  __ret; \
-})
-#else
-#define vld4_dup_s32(__p0) __extension__ ({ \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#else
-#define vld4_dup_s64(__p0) __extension__ ({ \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_s16(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
-  __ret; \
-})
-#else
-#define vld4_dup_s16(__p0) __extension__ ({ \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
-  __ret; \
-})
-#else
-#define vld4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  poly8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
-  __ret; \
-})
-#else
-#define vld4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  poly16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  poly16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
-  __ret; \
-})
-#else
-#define vld4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  poly16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
-  __ret; \
-})
-#else
-#define vld4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  uint32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  uint32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
-  __ret; \
-})
-#else
-#define vld4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  uint16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 41); \
-  __ret; \
-})
-#else
-#define vld4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  float32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  float32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 41); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 34); \
-  __ret; \
-})
-#else
-#define vld4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  int32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  int32x4x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 34); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 33); \
-  __ret; \
-})
-#else
-#define vld4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  int16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 33); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
-  __ret; \
-})
-#else
-#define vld4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  uint8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
-  __ret; \
-})
-#else
-#define vld4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  uint32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  uint32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
-  __ret; \
-})
-#else
-#define vld4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  uint16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  uint16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
-  __ret; \
-})
-#else
-#define vld4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  int8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 9); \
-  __ret; \
-})
-#else
-#define vld4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  float32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  float32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 9); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 2); \
-  __ret; \
-})
-#else
-#define vld4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  int32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  int32x2x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 2); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 1); \
-  __ret; \
-})
-#else
-#define vld4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  int16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  int16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 1); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vmax_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vmax_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vmax_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vmax_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vminq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vminq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vminq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vminq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vmin_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vmin_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vmin_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vmin_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint8x16_t vmlaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int8x16_t vmlaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float32x4_t vmlaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int32x4_t vmlaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int16x8_t vmlaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint8x8_t vmla_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint32x2_t vmla_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint16x4_t vmla_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int8x8_t vmla_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float32x2_t vmla_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int32x2_t vmla_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int16x4_t vmla_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x8_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlaq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlaq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __ret;
-  __ret = __p0 + __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai float32x4_t vmlaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 + __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int32x4_t vmlaq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int16x8_t vmlaq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint32x2_t __ret;
-  __ret = __p0 + __p1 * (uint32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai uint32x2_t vmla_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 + __rev1 * (uint32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint16x4_t __ret;
-  __ret = __p0 + __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint16x4_t vmla_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 + __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __ret;
-  __ret = __p0 + __p1 * (float32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai float32x2_t vmla_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 + __rev1 * (float32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int32x2_t __ret;
-  __ret = __p0 + __p1 * (int32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai int32x2_t vmla_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 + __rev1 * (int32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int16x4_t __ret;
-  __ret = __p0 + __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int16x4_t vmla_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 + __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint8x16_t vmlsq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlsq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int8x16_t vmlsq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float32x4_t vmlsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int16x8_t vmlsq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint8x8_t vmls_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint32x2_t vmls_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai uint16x4_t vmls_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int8x8_t vmls_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float32x2_t vmls_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int32x2_t vmls_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai int16x4_t vmls_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x8_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - __p1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsq_n_u32(uint32x4_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __rev1 * (uint32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 - __p1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlsq_n_u16(uint16x8_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __rev1 * (uint16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __ret;
-  __ret = __p0 - __p1 * (float32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai float32x4_t vmlsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 - __rev1 * (float32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - __p1 * (int32x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsq_n_s32(int32x4_t __p0, int32x4_t __p1, int32_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __rev1 * (int32x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 - __p1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int16x8_t vmlsq_n_s16(int16x8_t __p0, int16x8_t __p1, int16_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __rev1 * (int16x8_t) {__p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint32x2_t __ret;
-  __ret = __p0 - __p1 * (uint32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai uint32x2_t vmls_n_u32(uint32x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 - __rev1 * (uint32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint16x4_t __ret;
-  __ret = __p0 - __p1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai uint16x4_t vmls_n_u16(uint16x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 - __rev1 * (uint16x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __ret;
-  __ret = __p0 - __p1 * (float32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai float32x2_t vmls_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 - __rev1 * (float32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int32x2_t __ret;
-  __ret = __p0 - __p1 * (int32x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai int32x2_t vmls_n_s32(int32x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 - __rev1 * (int32x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int16x4_t __ret;
-  __ret = __p0 - __p1 * (int16x4_t) {__p2, __p2, __p2, __p2};
-  return __ret;
-}
-#else
-__ai int16x4_t vmls_n_s16(int16x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 - __rev1 * (int16x4_t) {__p2, __p2, __p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vmov_n_p8(poly8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly8x8_t vmov_n_p8(poly8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vmov_n_p16(poly16_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly16x4_t vmov_n_p16(poly16_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly8x16_t vmovq_n_p8(poly8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai poly16x8_t vmovq_n_p16(poly16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint8x16_t vmovq_n_u8(uint8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint32x4_t vmovq_n_u32(uint32_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai uint64x2_t vmovq_n_u64(uint64_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint16x8_t vmovq_n_u16(uint16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmovq_n_s8(int8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int8x16_t vmovq_n_s8(int8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmovq_n_f32(float32_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai float32x4_t vmovq_n_f32(float32_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmovq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vmovq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovq_n_s32(int32_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int32x4_t vmovq_n_s32(int32_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmovq_n_s64(int64_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai int64x2_t vmovq_n_s64(int64_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovq_n_s16(int16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int16x8_t vmovq_n_s16(int16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmov_n_u8(uint8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint8x8_t vmov_n_u8(uint8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmov_n_u32(uint32_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai uint32x2_t vmov_n_u32(uint32_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai uint64x1_t vmov_n_u64(uint64_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmov_n_u16(uint16_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai uint16x4_t vmov_n_u16(uint16_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmov_n_s8(int8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int8x8_t vmov_n_s8(int8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) {__p0, __p0, __p0, __p0, __p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmov_n_f32(float32_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai float32x2_t vmov_n_f32(float32_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmov_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vmov_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmov_n_s32(int32_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai int32x2_t vmov_n_s32(int32_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vmov_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai int64x1_t vmov_n_s64(int64_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmov_n_s16(int16_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
-  return __ret;
-}
-#else
-__ai int16x4_t vmov_n_s16(int16_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) {__p0, __p0, __p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vmovl_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vmovl_u8(uint8x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 49);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmovl_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmovl_u32(uint32x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 51);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmovl_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmovl_u16(uint16x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 50);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovl_s8(int8x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vmovl_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vmovl_s8(int8x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmovl_s32(int32x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vmovl_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmovl_s32(int32x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovl_s16(int16x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vmovl_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmovl_s16(int16x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmovl_v((int8x8_t)__p0, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vmovn_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vmovn_u32(uint32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vmovn_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vmovn_u64(uint64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vmovn_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vmovn_u16(uint16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmovn_s32(int32x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vmovn_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vmovn_s32(int32x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmovn_s64(int64x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vmovn_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vmovn_s64(int64x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmovn_s16(int16x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vmovn_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vmovn_s16(int16x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vmovn_v((int8x16_t)__p0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vmulq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vmulq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vmulq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vmulq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float32x4_t vmulq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vmulq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vmulq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vmul_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vmul_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vmul_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vmul_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float32x2_t vmul_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vmul_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vmul_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vmul_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vmul_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vmulq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vmulq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai uint32x4_t vmulq_n_u32(uint32x4_t __p0, uint32_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 * (uint32x4_t) {__p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai uint16x8_t vmulq_n_u16(uint16x8_t __p0, uint16_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 * (uint16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
-  float32x4_t __ret;
-  __ret = __p0 * (float32x4_t) {__p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai float32x4_t vmulq_n_f32(float32x4_t __p0, float32_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 * (float32x4_t) {__p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 * (int32x4_t) {__p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai int32x4_t vmulq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 * (int32x4_t) {__p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai int16x8_t vmulq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 * (int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 * (uint32x2_t) {__p1, __p1};
-  return __ret;
-}
-#else
-__ai uint32x2_t vmul_n_u32(uint32x2_t __p0, uint32_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 * (uint32x2_t) {__p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai uint16x4_t vmul_n_u16(uint16x4_t __p0, uint16_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 * (uint16x4_t) {__p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
-  float32x2_t __ret;
-  __ret = __p0 * (float32x2_t) {__p1, __p1};
-  return __ret;
-}
-#else
-__ai float32x2_t vmul_n_f32(float32x2_t __p0, float32_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 * (float32x2_t) {__p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 * (int32x2_t) {__p1, __p1};
-  return __ret;
-}
-#else
-__ai int32x2_t vmul_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 * (int32x2_t) {__p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 * (int16x4_t) {__p1, __p1, __p1, __p1};
-  return __ret;
-}
-#else
-__ai int16x4_t vmul_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 * (int16x4_t) {__p1, __p1, __p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
-  return __ret;
-}
-#else
-__ai poly16x8_t vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 37);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai poly16x8_t __noswap_vmull_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 37);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vmull_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 49);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmull_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 51);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmull_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 50);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vmull_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vmull_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = vmull_u32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __noswap_vmull_u32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = vmull_u16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __noswap_vmull_u16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmull_n_u32(uint32x2_t __p0, uint32_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint32x2_t) {__p1, __p1}, 51);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmull_n_u16(uint16x4_t __p0, uint16_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(uint16x4_t) {__p1, __p1, __p1, __p1}, 50);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai poly8x8_t vmvn_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai poly8x16_t vmvnq_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint8x16_t vmvnq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint32x4_t vmvnq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint16x8_t vmvnq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmvnq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int8x16_t vmvnq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmvnq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int32x4_t vmvnq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmvnq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int16x8_t vmvnq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint8x8_t vmvn_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint32x2_t vmvn_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai uint16x4_t vmvn_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vmvn_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int8x8_t vmvn_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vmvn_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int32x2_t vmvn_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vmvn_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = ~__p0;
-  return __ret;
-}
-#else
-__ai int16x4_t vmvn_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = ~__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vnegq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int8x16_t vnegq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vnegq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float32x4_t vnegq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vnegq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int32x4_t vnegq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vnegq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int16x8_t vnegq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vneg_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int8x8_t vneg_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vneg_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float32x2_t vneg_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vneg_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int32x2_t vneg_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vneg_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int16x4_t vneg_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vornq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vornq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vornq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vornq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vornq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vornq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vornq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vornq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vorn_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vorn_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vorn_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vorn_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vorn_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vorn_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vorn_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 | ~__p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vorn_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 | ~__rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vorrq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vorrq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vorrq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vorrq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vorrq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vorrq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vorrq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vorrq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vorr_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vorr_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vorr_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vorr_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vorr_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vorr_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vorr_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 | __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vorr_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 | __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vpadalq_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vpadalq_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vpadalq_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vpadalq_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vpadalq_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vpadalq_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpadalq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vpadal_u8(uint16x4_t __p0, uint8x8_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vpadal_u32(uint64x1_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vpadal_u16(uint32x2_t __p0, uint16x4_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vpadal_s8(int16x4_t __p0, int8x8_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vpadal_s32(int64x1_t __p0, int32x2_t __p1) {
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__rev1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vpadal_s16(int32x2_t __p0, int16x4_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpadal_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vpadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vpadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vpadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vpadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vpadd_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vpadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vpadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vpaddlq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vpaddlq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vpaddlq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vpaddlq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vpaddlq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vpaddlq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpaddlq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vpaddl_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vpaddl_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vpaddl_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vpaddl_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vpaddl_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vpaddl_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vpaddl_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vpaddl_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vpaddl_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpaddl_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vpmax_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vpmax_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vpmax_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vpmax_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vpmax_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vpmax_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vpmax_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vpmin_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vpmin_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vpmin_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vpmin_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vpmin_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vpmin_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vpmin_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqabsq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqabsq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqabsq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqabsq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqabsq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqabsq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqabs_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqabs_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqabs_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqabs_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqabs_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqabs_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqabs_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vqaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vqaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vqaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vqaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vqadd_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vqadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqadd_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vqadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlal_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlal_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlal_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlal_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlal_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlsl_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlsl_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlsl_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlsl_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int32x2_t) {__p2, __p2}, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__rev0, (int8x8_t)__rev1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmlsl_v((int8x16_t)__p0, (int8x8_t)__p1, (int8x8_t)(int16x4_t) {__p2, __p2, __p2, __p2}, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vqdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vqdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vqdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = vqdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = vqdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = vqdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__p0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqdmulh_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmull_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmull_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vqdmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vqdmull_n_s32(int32x2_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 35);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqdmull_n_s16(int16x4_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqdmull_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqmovn_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vqmovn_u32(uint32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqmovn_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vqmovn_u64(uint64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqmovn_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vqmovn_u16(uint16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqmovn_s32(int32x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqmovn_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vqmovn_s32(int32x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqmovn_s64(int64x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqmovn_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vqmovn_s64(int64x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqmovn_s16(int16x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqmovn_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vqmovn_s16(int16x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqmovn_v((int8x16_t)__p0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqmovun_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vqmovun_s32(int32x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqmovun_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vqmovun_s64(int64x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqmovun_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vqmovun_s16(int16x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqmovun_v((int8x16_t)__p0, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqnegq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqnegq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqnegq_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqnegq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqnegq_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqnegq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqneg_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqneg_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqneg_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqneg_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqneg_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqneg_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqneg_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqrdmulhq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vqrdmulhq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vqrdmulh_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vqrdmulh_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqrdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulhq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqrdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = vqrdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulhq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqrdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = vqrdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulh_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqrdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = vqrdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulh_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqrdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqrdmulhq_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int32x4_t) {__p1, __p1, __p1, __p1}, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__p0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqrdmulhq_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrdmulhq_v((int8x16_t)__rev0, (int8x16_t)(int16x8_t) {__p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1}, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqrdmulh_n_s32(int32x2_t __p0, int32_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)(int32x2_t) {__p1, __p1}, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__p0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqrdmulh_n_s16(int16x4_t __p0, int16_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrdmulh_v((int8x8_t)__rev0, (int8x8_t)(int16x4_t) {__p1, __p1, __p1, __p1}, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vqrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vqrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vqrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vqrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqrshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqrshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqrshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqrshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqrshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqrshrun_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vqshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vqshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vqshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vqshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 48); \
-  __ret; \
-})
-#else
-#define vqshlq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vqshlq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vqshlq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vqshlq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 32); \
-  __ret; \
-})
-#else
-#define vqshlq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vqshlq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vqshlq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vqshlq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vqshlq_n_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqshl_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqshl_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vqshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqshl_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vqshl_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vqshl_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vqshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshl_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vqshl_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqshl_n_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 48); \
-  __ret; \
-})
-#else
-#define vqshluq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vqshluq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vqshluq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vqshluq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vqshluq_n_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqshlu_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqshlu_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vqshlu_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqshlu_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshlu_n_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vqshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vqshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vqshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vqshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vqshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrun_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vqshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrun_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vqshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vqshrun_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vqshrun_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vqsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vqsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vqsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vqsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqsubq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vqsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vqsubq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vqsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vqsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vqsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vqsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vqsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vqsub_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vraddhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vraddhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vraddhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vraddhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vraddhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vraddhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vraddhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vrecpeq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrecpeq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrecpe_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrecpe_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrecpe_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrecpsq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrecps_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrecps_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai poly8x8_t vrev16_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  return __ret;
-}
-#else
-__ai poly8x16_t vrev16q_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrev16q_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrev16q_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  return __ret;
-}
-#else
-__ai int8x16_t vrev16q_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrev16_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrev16_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai int8x8_t vrev16_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vrev32_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai poly16x4_t vrev32_p16(poly16x4_t __p0) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  return __ret;
-}
-#else
-__ai poly8x16_t vrev32q_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai poly16x8_t vrev32q_p16(poly16x8_t __p0) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrev32q_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai uint16x8_t vrev32q_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrev32q_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  return __ret;
-}
-#else
-__ai int8x16_t vrev32q_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vrev32q_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2, 5, 4, 7, 6);
-  return __ret;
-}
-#else
-__ai int16x8_t vrev32q_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2, 5, 4, 7, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrev32_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai uint16x4_t vrev32_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrev32_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai int8x8_t vrev32_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vrev32_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai int16x4_t vrev32_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vrev64_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vrev64_p16(poly16x4_t __p0) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  return __ret;
-}
-#else
-__ai poly8x16_t vrev64q_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai poly16x8_t vrev64q_p16(poly16x8_t __p0) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrev64q_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vrev64q_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai uint16x8_t vrev64q_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrev64q_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  return __ret;
-}
-#else
-__ai int8x16_t vrev64q_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrev64q_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai float32x4_t vrev64q_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vrev64q_s32(int32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0, 3, 2);
-  return __ret;
-}
-#else
-__ai int32x4_t vrev64q_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0, 3, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vrev64q_s16(int16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai int16x8_t vrev64q_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrev64_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrev64_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vrev64_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrev64_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vrev64_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrev64_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
-  return __ret;
-}
-#else
-__ai float32x2_t vrev64_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vrev64_s32(int32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1, 0);
-  return __ret;
-}
-#else
-__ai int32x2_t vrev64_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vrev64_s16(int16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai int16x4_t vrev64_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrhaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vrhaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vrhaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vrhaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vrhaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vrhaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vrhaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrhadd_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrhadd_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vrhadd_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vrhadd_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vrhadd_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vrhadd_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrhadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vrshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vrshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vrshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vrshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vrshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vrshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vrshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vrshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vrshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vrshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vrshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vrshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vrshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vrshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 48); \
-  __ret; \
-})
-#else
-#define vrshrq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vrshrq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vrshrq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vrshrq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 32); \
-  __ret; \
-})
-#else
-#define vrshrq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vrshrq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vrshrq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vrshrq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vrshrq_n_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vrshr_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vrshr_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vrshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vrshr_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vrshr_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vrshr_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vrshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshr_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vrshr_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrshr_n_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vrshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vrsqrteq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrsqrteq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrsqrte_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrsqrte_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrsqrtsq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrsqrts_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrsqrts_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vrsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vrsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vrsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vrsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vrsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vrsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vrsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vrsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vrsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vrsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vrsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vrsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vrsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vrsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vrsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vrsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vrsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vrsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vrsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vrsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vrsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vrsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vrsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vrsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsetq_lane_i8(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vsetq_lane_f32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsetq_lane_i32(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsetq_lane_i16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vset_lane_i8(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vset_lane_f32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vset_lane_i32(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vset_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vset_lane_i16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vshlq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vshlq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vshlq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vshlq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vshlq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vshlq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vshlq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vshlq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vshlq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vshl_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vshl_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vshl_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vshl_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vshl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vshl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vshl_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vshl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vshl_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 48); \
-  __ret; \
-})
-#else
-#define vshlq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vshlq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vshlq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vshlq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 32); \
-  __ret; \
-})
-#else
-#define vshlq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vshlq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vshlq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshlq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vshlq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshlq_n_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vshl_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vshl_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vshl_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vshl_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vshl_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vshl_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vshl_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshl_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vshl_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshl_n_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vshll_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 49); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vshll_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 51); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vshll_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 50); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vshll_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 33); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vshll_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 35); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vshll_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshll_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshll_n_v((int8x8_t)__s0, __p1, 34); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 48); \
-  __ret; \
-})
-#else
-#define vshrq_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 50); \
-  __ret; \
-})
-#else
-#define vshrq_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vshrq_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vshrq_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 32); \
-  __ret; \
-})
-#else
-#define vshrq_n_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 34); \
-  __ret; \
-})
-#else
-#define vshrq_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vshrq_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vshrq_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vshrq_n_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vshr_n_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vshr_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vshr_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vshr_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vshr_n_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vshr_n_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vshr_n_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshr_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vshr_n_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshr_n_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 17); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#else
-#define vshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 18); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#else
-#define vshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 16); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#else
-#define vshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#else
-#define vshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__rev0, __p1, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vshrn_n_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vshrn_n_v((int8x16_t)__s0, __p1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
-  __ret; \
-})
-#else
-#define vsli_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
-  __ret; \
-})
-#else
-#define vsli_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
-  __ret; \
-})
-#else
-#define vsliq_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
-  __ret; \
-})
-#else
-#define vsliq_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vsliq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vsliq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vsliq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vsliq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vsliq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vsliq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vsliq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vsliq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vsli_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vsli_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vsli_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vsli_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vsli_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vsli_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vsli_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vsli_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsli_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vsraq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vsraq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vsraq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vsraq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vsraq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vsraq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vsraq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vsraq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsraq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vsra_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vsra_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vsra_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vsra_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vsra_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vsra_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vsra_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vsra_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsra_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 4); \
-  __ret; \
-})
-#else
-#define vsri_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = (poly8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 4); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 5); \
-  __ret; \
-})
-#else
-#define vsri_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = (poly16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 5); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 36); \
-  __ret; \
-})
-#else
-#define vsriq_n_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = (poly8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 36); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 37); \
-  __ret; \
-})
-#else
-#define vsriq_n_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = (poly16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 37); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 48); \
-  __ret; \
-})
-#else
-#define vsriq_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = (uint8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 48); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 50); \
-  __ret; \
-})
-#else
-#define vsriq_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = (uint32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 50); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 51); \
-  __ret; \
-})
-#else
-#define vsriq_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 49); \
-  __ret; \
-})
-#else
-#define vsriq_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 32); \
-  __ret; \
-})
-#else
-#define vsriq_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = (int8x16_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 32); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 34); \
-  __ret; \
-})
-#else
-#define vsriq_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = (int32x4_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 34); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 35); \
-  __ret; \
-})
-#else
-#define vsriq_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 33); \
-  __ret; \
-})
-#else
-#define vsriq_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 16); \
-  __ret; \
-})
-#else
-#define vsri_n_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = (uint8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 16); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 18); \
-  __ret; \
-})
-#else
-#define vsri_n_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = (uint32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 18); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#else
-#define vsri_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64x1_t __s1 = __p1; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 17); \
-  __ret; \
-})
-#else
-#define vsri_n_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 0); \
-  __ret; \
-})
-#else
-#define vsri_n_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = (int8x8_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 0); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 2); \
-  __ret; \
-})
-#else
-#define vsri_n_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __ret; \
-  __ret = (int32x2_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#else
-#define vsri_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64x1_t __s1 = __p1; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 1); \
-  __ret; \
-})
-#else
-#define vsri_n_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vsri_n_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 4); \
-})
-#else
-#define vst1_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 5); \
-})
-#else
-#define vst1_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 36); \
-})
-#else
-#define vst1q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 37); \
-})
-#else
-#define vst1q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 48); \
-})
-#else
-#define vst1q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 50); \
-})
-#else
-#define vst1q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 51); \
-})
-#else
-#define vst1q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 49); \
-})
-#else
-#define vst1q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 32); \
-})
-#else
-#define vst1q_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 41); \
-})
-#else
-#define vst1q_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 34); \
-})
-#else
-#define vst1q_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 35); \
-})
-#else
-#define vst1q_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 33); \
-})
-#else
-#define vst1q_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 16); \
-})
-#else
-#define vst1_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 18); \
-})
-#else
-#define vst1_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
-})
-#else
-#define vst1_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 17); \
-})
-#else
-#define vst1_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 0); \
-})
-#else
-#define vst1_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 9); \
-})
-#else
-#define vst1_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 2); \
-})
-#else
-#define vst1_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
-})
-#else
-#define vst1_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 1); \
-})
-#else
-#define vst1_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 4); \
-})
-#else
-#define vst1_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8_t __s1 = __p1; \
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 5); \
-})
-#else
-#define vst1_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4_t __s1 = __p1; \
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 36); \
-})
-#else
-#define vst1q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16_t __s1 = __p1; \
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 37); \
-})
-#else
-#define vst1q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8_t __s1 = __p1; \
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 48); \
-})
-#else
-#define vst1q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 50); \
-})
-#else
-#define vst1q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 51); \
-})
-#else
-#define vst1q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2_t __s1 = __p1; \
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 49); \
-})
-#else
-#define vst1q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 32); \
-})
-#else
-#define vst1q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 41); \
-})
-#else
-#define vst1q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 34); \
-})
-#else
-#define vst1q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 35); \
-})
-#else
-#define vst1q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2_t __s1 = __p1; \
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 33); \
-})
-#else
-#define vst1q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 16); \
-})
-#else
-#define vst1_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 18); \
-})
-#else
-#define vst1_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-})
-#else
-#define vst1_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 17); \
-})
-#else
-#define vst1_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 0); \
-})
-#else
-#define vst1_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 9); \
-})
-#else
-#define vst1_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 2); \
-})
-#else
-#define vst1_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-})
-#else
-#define vst1_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 1); \
-})
-#else
-#define vst1_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p8_x2(__p0, __p1) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
-})
-#else
-#define vst1_p8_x2(__p0, __p1) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  poly8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p16_x2(__p0, __p1) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
-})
-#else
-#define vst1_p16_x2(__p0, __p1) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  poly16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
-})
-#else
-#define vst1q_p8_x2(__p0, __p1) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  poly8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
-})
-#else
-#define vst1q_p16_x2(__p0, __p1) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  poly16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
-})
-#else
-#define vst1q_u8_x2(__p0, __p1) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  uint8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
-})
-#else
-#define vst1q_u32_x2(__p0, __p1) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  uint32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
-})
-#else
-#define vst1q_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  uint64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
-})
-#else
-#define vst1q_u16_x2(__p0, __p1) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  uint16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
-})
-#else
-#define vst1q_s8_x2(__p0, __p1) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  int8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 41); \
-})
-#else
-#define vst1q_f32_x2(__p0, __p1) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  float32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 34); \
-})
-#else
-#define vst1q_s32_x2(__p0, __p1) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  int32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 35); \
-})
-#else
-#define vst1q_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  int64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 33); \
-})
-#else
-#define vst1q_s16_x2(__p0, __p1) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  int16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u8_x2(__p0, __p1) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
-})
-#else
-#define vst1_u8_x2(__p0, __p1) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  uint8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u32_x2(__p0, __p1) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
-})
-#else
-#define vst1_u32_x2(__p0, __p1) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  uint32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#else
-#define vst1_u64_x2(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u16_x2(__p0, __p1) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
-})
-#else
-#define vst1_u16_x2(__p0, __p1) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  uint16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s8_x2(__p0, __p1) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
-})
-#else
-#define vst1_s8_x2(__p0, __p1) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  int8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f32_x2(__p0, __p1) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 9); \
-})
-#else
-#define vst1_f32_x2(__p0, __p1) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  float32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s32_x2(__p0, __p1) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 2); \
-})
-#else
-#define vst1_s32_x2(__p0, __p1) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  int32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#else
-#define vst1_s64_x2(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s16_x2(__p0, __p1) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 1); \
-})
-#else
-#define vst1_s16_x2(__p0, __p1) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  int16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p8_x3(__p0, __p1) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
-})
-#else
-#define vst1_p8_x3(__p0, __p1) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  poly8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p16_x3(__p0, __p1) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
-})
-#else
-#define vst1_p16_x3(__p0, __p1) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  poly16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
-})
-#else
-#define vst1q_p8_x3(__p0, __p1) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  poly8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
-})
-#else
-#define vst1q_p16_x3(__p0, __p1) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  poly16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
-})
-#else
-#define vst1q_u8_x3(__p0, __p1) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  uint8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
-})
-#else
-#define vst1q_u32_x3(__p0, __p1) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  uint32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
-})
-#else
-#define vst1q_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  uint64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
-})
-#else
-#define vst1q_u16_x3(__p0, __p1) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  uint16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
-})
-#else
-#define vst1q_s8_x3(__p0, __p1) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  int8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 41); \
-})
-#else
-#define vst1q_f32_x3(__p0, __p1) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  float32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 34); \
-})
-#else
-#define vst1q_s32_x3(__p0, __p1) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  int32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 35); \
-})
-#else
-#define vst1q_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  int64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 33); \
-})
-#else
-#define vst1q_s16_x3(__p0, __p1) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  int16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u8_x3(__p0, __p1) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
-})
-#else
-#define vst1_u8_x3(__p0, __p1) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  uint8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u32_x3(__p0, __p1) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
-})
-#else
-#define vst1_u32_x3(__p0, __p1) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  uint32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#else
-#define vst1_u64_x3(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u16_x3(__p0, __p1) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
-})
-#else
-#define vst1_u16_x3(__p0, __p1) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  uint16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s8_x3(__p0, __p1) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
-})
-#else
-#define vst1_s8_x3(__p0, __p1) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  int8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f32_x3(__p0, __p1) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 9); \
-})
-#else
-#define vst1_f32_x3(__p0, __p1) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  float32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s32_x3(__p0, __p1) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 2); \
-})
-#else
-#define vst1_s32_x3(__p0, __p1) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  int32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#else
-#define vst1_s64_x3(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s16_x3(__p0, __p1) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 1); \
-})
-#else
-#define vst1_s16_x3(__p0, __p1) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  int16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p8_x4(__p0, __p1) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
-})
-#else
-#define vst1_p8_x4(__p0, __p1) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  poly8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p16_x4(__p0, __p1) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
-})
-#else
-#define vst1_p16_x4(__p0, __p1) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  poly16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
-})
-#else
-#define vst1q_p8_x4(__p0, __p1) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  poly8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
-})
-#else
-#define vst1q_p16_x4(__p0, __p1) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  poly16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
-})
-#else
-#define vst1q_u8_x4(__p0, __p1) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  uint8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
-})
-#else
-#define vst1q_u32_x4(__p0, __p1) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  uint32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
-})
-#else
-#define vst1q_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  uint64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
-})
-#else
-#define vst1q_u16_x4(__p0, __p1) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  uint16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
-})
-#else
-#define vst1q_s8_x4(__p0, __p1) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  int8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 41); \
-})
-#else
-#define vst1q_f32_x4(__p0, __p1) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  float32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 34); \
-})
-#else
-#define vst1q_s32_x4(__p0, __p1) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  int32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 35); \
-})
-#else
-#define vst1q_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  int64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 33); \
-})
-#else
-#define vst1q_s16_x4(__p0, __p1) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  int16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u8_x4(__p0, __p1) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
-})
-#else
-#define vst1_u8_x4(__p0, __p1) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  uint8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u32_x4(__p0, __p1) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
-})
-#else
-#define vst1_u32_x4(__p0, __p1) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  uint32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#else
-#define vst1_u64_x4(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_u16_x4(__p0, __p1) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
-})
-#else
-#define vst1_u16_x4(__p0, __p1) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  uint16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s8_x4(__p0, __p1) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
-})
-#else
-#define vst1_s8_x4(__p0, __p1) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  int8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f32_x4(__p0, __p1) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 9); \
-})
-#else
-#define vst1_f32_x4(__p0, __p1) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  float32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s32_x4(__p0, __p1) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 2); \
-})
-#else
-#define vst1_s32_x4(__p0, __p1) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  int32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#else
-#define vst1_s64_x4(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_s16_x4(__p0, __p1) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 1); \
-})
-#else
-#define vst1_s16_x4(__p0, __p1) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  int16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 4); \
-})
-#else
-#define vst2_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  poly8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 5); \
-})
-#else
-#define vst2_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  poly16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 36); \
-})
-#else
-#define vst2q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  poly8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 37); \
-})
-#else
-#define vst2q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  poly16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 48); \
-})
-#else
-#define vst2q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  uint8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 50); \
-})
-#else
-#define vst2q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  uint32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 49); \
-})
-#else
-#define vst2q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  uint16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 32); \
-})
-#else
-#define vst2q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  int8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 41); \
-})
-#else
-#define vst2q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  float32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 34); \
-})
-#else
-#define vst2q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  int32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 33); \
-})
-#else
-#define vst2q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  int16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 16); \
-})
-#else
-#define vst2_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  uint8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 18); \
-})
-#else
-#define vst2_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  uint32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#else
-#define vst2_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 17); \
-})
-#else
-#define vst2_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  uint16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_s8(__p0, __p1) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 0); \
-})
-#else
-#define vst2_s8(__p0, __p1) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  int8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_f32(__p0, __p1) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 9); \
-})
-#else
-#define vst2_f32(__p0, __p1) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  float32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_s32(__p0, __p1) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 2); \
-})
-#else
-#define vst2_s32(__p0, __p1) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  int32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_s64(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#else
-#define vst2_s64(__p0, __p1) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_s16(__p0, __p1) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 1); \
-})
-#else
-#define vst2_s16(__p0, __p1) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  int16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 4); \
-})
-#else
-#define vst2_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x2_t __s1 = __p1; \
-  poly8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 5); \
-})
-#else
-#define vst2_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x2_t __s1 = __p1; \
-  poly16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 37); \
-})
-#else
-#define vst2q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x2_t __s1 = __p1; \
-  poly16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 50); \
-})
-#else
-#define vst2q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x2_t __s1 = __p1; \
-  uint32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 49); \
-})
-#else
-#define vst2q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x2_t __s1 = __p1; \
-  uint16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 41); \
-})
-#else
-#define vst2q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x2_t __s1 = __p1; \
-  float32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 34); \
-})
-#else
-#define vst2q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x2_t __s1 = __p1; \
-  int32x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 33); \
-})
-#else
-#define vst2q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x2_t __s1 = __p1; \
-  int16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 16); \
-})
-#else
-#define vst2_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x2_t __s1 = __p1; \
-  uint8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 18); \
-})
-#else
-#define vst2_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x2_t __s1 = __p1; \
-  uint32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 17); \
-})
-#else
-#define vst2_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x2_t __s1 = __p1; \
-  uint16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 0); \
-})
-#else
-#define vst2_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x2_t __s1 = __p1; \
-  int8x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], __p2, 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 9); \
-})
-#else
-#define vst2_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x2_t __s1 = __p1; \
-  float32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 2); \
-})
-#else
-#define vst2_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x2_t __s1 = __p1; \
-  int32x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 1); \
-})
-#else
-#define vst2_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x2_t __s1 = __p1; \
-  int16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 4); \
-})
-#else
-#define vst3_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  poly8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 5); \
-})
-#else
-#define vst3_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  poly16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 36); \
-})
-#else
-#define vst3q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  poly8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 37); \
-})
-#else
-#define vst3q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  poly16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 48); \
-})
-#else
-#define vst3q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  uint8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 50); \
-})
-#else
-#define vst3q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  uint32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 49); \
-})
-#else
-#define vst3q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  uint16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 32); \
-})
-#else
-#define vst3q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  int8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 41); \
-})
-#else
-#define vst3q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  float32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 34); \
-})
-#else
-#define vst3q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  int32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 33); \
-})
-#else
-#define vst3q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  int16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 16); \
-})
-#else
-#define vst3_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  uint8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 18); \
-})
-#else
-#define vst3_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  uint32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#else
-#define vst3_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 17); \
-})
-#else
-#define vst3_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  uint16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_s8(__p0, __p1) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 0); \
-})
-#else
-#define vst3_s8(__p0, __p1) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  int8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_f32(__p0, __p1) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 9); \
-})
-#else
-#define vst3_f32(__p0, __p1) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  float32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_s32(__p0, __p1) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 2); \
-})
-#else
-#define vst3_s32(__p0, __p1) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  int32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_s64(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#else
-#define vst3_s64(__p0, __p1) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_s16(__p0, __p1) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 1); \
-})
-#else
-#define vst3_s16(__p0, __p1) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  int16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 4); \
-})
-#else
-#define vst3_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x3_t __s1 = __p1; \
-  poly8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 5); \
-})
-#else
-#define vst3_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x3_t __s1 = __p1; \
-  poly16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 37); \
-})
-#else
-#define vst3q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x3_t __s1 = __p1; \
-  poly16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 50); \
-})
-#else
-#define vst3q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x3_t __s1 = __p1; \
-  uint32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 49); \
-})
-#else
-#define vst3q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x3_t __s1 = __p1; \
-  uint16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 41); \
-})
-#else
-#define vst3q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x3_t __s1 = __p1; \
-  float32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 34); \
-})
-#else
-#define vst3q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x3_t __s1 = __p1; \
-  int32x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 33); \
-})
-#else
-#define vst3q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x3_t __s1 = __p1; \
-  int16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 16); \
-})
-#else
-#define vst3_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x3_t __s1 = __p1; \
-  uint8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 18); \
-})
-#else
-#define vst3_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x3_t __s1 = __p1; \
-  uint32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 17); \
-})
-#else
-#define vst3_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x3_t __s1 = __p1; \
-  uint16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 0); \
-})
-#else
-#define vst3_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x3_t __s1 = __p1; \
-  int8x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], __p2, 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 9); \
-})
-#else
-#define vst3_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x3_t __s1 = __p1; \
-  float32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 2); \
-})
-#else
-#define vst3_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x3_t __s1 = __p1; \
-  int32x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 1); \
-})
-#else
-#define vst3_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x3_t __s1 = __p1; \
-  int16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 4); \
-})
-#else
-#define vst4_p8(__p0, __p1) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  poly8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 5); \
-})
-#else
-#define vst4_p16(__p0, __p1) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  poly16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 36); \
-})
-#else
-#define vst4q_p8(__p0, __p1) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  poly8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 37); \
-})
-#else
-#define vst4q_p16(__p0, __p1) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  poly16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 48); \
-})
-#else
-#define vst4q_u8(__p0, __p1) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  uint8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 50); \
-})
-#else
-#define vst4q_u32(__p0, __p1) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  uint32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 49); \
-})
-#else
-#define vst4q_u16(__p0, __p1) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  uint16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 32); \
-})
-#else
-#define vst4q_s8(__p0, __p1) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  int8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 41); \
-})
-#else
-#define vst4q_f32(__p0, __p1) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  float32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 34); \
-})
-#else
-#define vst4q_s32(__p0, __p1) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  int32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 33); \
-})
-#else
-#define vst4q_s16(__p0, __p1) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  int16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 16); \
-})
-#else
-#define vst4_u8(__p0, __p1) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  uint8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 18); \
-})
-#else
-#define vst4_u32(__p0, __p1) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  uint32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#else
-#define vst4_u64(__p0, __p1) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 17); \
-})
-#else
-#define vst4_u16(__p0, __p1) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  uint16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_s8(__p0, __p1) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 0); \
-})
-#else
-#define vst4_s8(__p0, __p1) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  int8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_f32(__p0, __p1) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 9); \
-})
-#else
-#define vst4_f32(__p0, __p1) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  float32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_s32(__p0, __p1) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 2); \
-})
-#else
-#define vst4_s32(__p0, __p1) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  int32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_s64(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#else
-#define vst4_s64(__p0, __p1) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_s16(__p0, __p1) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 1); \
-})
-#else
-#define vst4_s16(__p0, __p1) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  int16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 4); \
-})
-#else
-#define vst4_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x8x4_t __s1 = __p1; \
-  poly8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 4); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 5); \
-})
-#else
-#define vst4_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x4x4_t __s1 = __p1; \
-  poly16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 5); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 37); \
-})
-#else
-#define vst4q_lane_p16(__p0, __p1, __p2) __extension__ ({ \
-  poly16x8x4_t __s1 = __p1; \
-  poly16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 37); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 50); \
-})
-#else
-#define vst4q_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4x4_t __s1 = __p1; \
-  uint32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 50); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 49); \
-})
-#else
-#define vst4q_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8x4_t __s1 = __p1; \
-  uint16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 49); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 41); \
-})
-#else
-#define vst4q_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4x4_t __s1 = __p1; \
-  float32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 41); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 34); \
-})
-#else
-#define vst4q_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4x4_t __s1 = __p1; \
-  int32x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 34); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 33); \
-})
-#else
-#define vst4q_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8x4_t __s1 = __p1; \
-  int16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 33); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 16); \
-})
-#else
-#define vst4_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x8x4_t __s1 = __p1; \
-  uint8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 16); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 18); \
-})
-#else
-#define vst4_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2x4_t __s1 = __p1; \
-  uint32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 18); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 17); \
-})
-#else
-#define vst4_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4x4_t __s1 = __p1; \
-  uint16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 17); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 0); \
-})
-#else
-#define vst4_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x8x4_t __s1 = __p1; \
-  int8x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], __p2, 0); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 9); \
-})
-#else
-#define vst4_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2x4_t __s1 = __p1; \
-  float32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 9); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 2); \
-})
-#else
-#define vst4_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2x4_t __s1 = __p1; \
-  int32x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 2); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 1); \
-})
-#else
-#define vst4_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4x4_t __s1 = __p1; \
-  int16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint8x16_t vsubq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint64x2_t vsubq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int8x16_t vsubq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float32x4_t vsubq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int32x4_t vsubq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int64x2_t vsubq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int16x8_t vsubq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint8x8_t vsub_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint32x2_t vsub_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint64x1_t vsub_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai uint16x4_t vsub_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int8x8_t vsub_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float32x2_t vsub_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int32x2_t vsub_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int64x1_t vsub_s64(int64x1_t __p0, int64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai int16x4_t vsub_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x4_t __noswap_vsubhn_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 17);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vsubhn_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint8x8_t __noswap_vsubhn_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 16);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x4_t __noswap_vsubhn_s32(int32x4_t __p0, int32x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vsubhn_s64(int64x2_t __p0, int64x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__rev0, (int8x16_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int8x8_t __noswap_vsubhn_s16(int16x8_t __p0, int16x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vsubhn_v((int8x16_t)__p0, (int8x16_t)__p1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = vmovl_u8(__p0) - vmovl_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmovl_u8(__rev0) - __noswap_vmovl_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmovl_u32(__p0) - vmovl_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vsubl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmovl_u32(__rev0) - __noswap_vmovl_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmovl_u16(__p0) - vmovl_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmovl_u16(__rev0) - __noswap_vmovl_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = vmovl_s8(__p0) - vmovl_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vsubl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmovl_s8(__rev0) - __noswap_vmovl_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = vmovl_s32(__p0) - vmovl_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vsubl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmovl_s32(__rev0) - __noswap_vmovl_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = vmovl_s16(__p0) - vmovl_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vsubl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmovl_s16(__rev0) - __noswap_vmovl_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 - vmovl_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubw_u8(uint16x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __noswap_vmovl_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 - vmovl_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vsubw_u32(uint64x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 - __noswap_vmovl_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 - vmovl_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubw_u16(uint32x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __noswap_vmovl_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 - vmovl_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vsubw_s8(int16x8_t __p0, int8x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __noswap_vmovl_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 - vmovl_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vsubw_s32(int64x2_t __p0, int32x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 - __noswap_vmovl_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 - vmovl_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vsubw_s16(int32x4_t __p0, int16x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __noswap_vmovl_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbl1_p8(poly8x8_t __p0, uint8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbl1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbl1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl1_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbl2_p8(poly8x8x2_t __p0, uint8x8_t __p1) {
-  poly8x8x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbl2_u8(uint8x8x2_t __p0, uint8x8_t __p1) {
-  uint8x8x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbl2_s8(int8x8x2_t __p0, int8x8_t __p1) {
-  int8x8x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl2_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbl3_p8(poly8x8x3_t __p0, uint8x8_t __p1) {
-  poly8x8x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbl3_u8(uint8x8x3_t __p0, uint8x8_t __p1) {
-  uint8x8x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbl3_s8(int8x8x3_t __p0, int8x8_t __p1) {
-  int8x8x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl3_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbl4_p8(poly8x8x4_t __p0, uint8x8_t __p1) {
-  poly8x8x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbl4_u8(uint8x8x4_t __p0, uint8x8_t __p1) {
-  uint8x8x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__p0.val[0], (int8x8_t)__p0.val[1], (int8x8_t)__p0.val[2], (int8x8_t)__p0.val[3], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbl4_s8(int8x8x4_t __p0, int8x8_t __p1) {
-  int8x8x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbl4_v((int8x8_t)__rev0.val[0], (int8x8_t)__rev0.val[1], (int8x8_t)__rev0.val[2], (int8x8_t)__rev0.val[3], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbx1_p8(poly8x8_t __p0, poly8x8_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbx1_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbx1_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx1_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbx2_p8(poly8x8_t __p0, poly8x8x2_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbx2_u8(uint8x8_t __p0, uint8x8x2_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbx2_s8(int8x8_t __p0, int8x8x2_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx2_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbx3_p8(poly8x8_t __p0, poly8x8x3_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbx3_u8(uint8x8_t __p0, uint8x8x3_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbx3_s8(int8x8_t __p0, int8x8x3_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx3_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtbx4_p8(poly8x8_t __p0, poly8x8x4_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtbx4_u8(uint8x8_t __p0, uint8x8x4_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__p0, (int8x8_t)__p1.val[0], (int8x8_t)__p1.val[1], (int8x8_t)__p1.val[2], (int8x8_t)__p1.val[3], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vtbx4_s8(int8x8_t __p0, int8x8x4_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vtbx4_v((int8x8_t)__rev0, (int8x8_t)__rev1.val[0], (int8x8_t)__rev1.val[1], (int8x8_t)__rev1.val[2], (int8x8_t)__rev1.val[3], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8x2_t vtrn_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
-  return __ret;
-}
-#else
-__ai poly16x4x2_t vtrn_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16x2_t vtrnq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
-  return __ret;
-}
-#else
-__ai poly16x8x2_t vtrnq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16x2_t vtrnq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4x2_t vtrnq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8x2_t vtrnq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16x2_t vtrnq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4x2_t vtrnq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4x2_t vtrnq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8x2_t vtrnq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8x2_t vtrn_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2x2_t vtrn_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4x2_t vtrn_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8x2_t vtrn_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2x2_t vtrn_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2x2_t vtrn_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4x2_t vtrn_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtst_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vtst_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vtstq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vtstq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vtstq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vtstq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vtstq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vtstq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vtstq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vtstq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtst_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vtst_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vtst_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtst_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vtst_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vtst_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vtst_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8x2_t vuzp_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
-  return __ret;
-}
-#else
-__ai poly16x4x2_t vuzp_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16x2_t vuzpq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
-  return __ret;
-}
-#else
-__ai poly16x8x2_t vuzpq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16x2_t vuzpq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4x2_t vuzpq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8x2_t vuzpq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16x2_t vuzpq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4x2_t vuzpq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4x2_t vuzpq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8x2_t vuzpq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8x2_t vuzp_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2x2_t vuzp_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4x2_t vuzp_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8x2_t vuzp_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2x2_t vuzp_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2x2_t vuzp_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4x2_t vuzp_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8x2_t vzip_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 4);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 5);
-  return __ret;
-}
-#else
-__ai poly16x4x2_t vzip_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 5);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16x2_t vzipq_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 36);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 37);
-  return __ret;
-}
-#else
-__ai poly16x8x2_t vzipq_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 37);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16x2_t vzipq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4x2_t vzipq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8x2_t vzipq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16x2_t vzipq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4x2_t vzipq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4x2_t vzipq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8x2_t vzipq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8x2_t vzip_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2x2_t vzip_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4x2_t vzip_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8x2_t vzip_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2x2_t vzip_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2x2_t vzip_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4x2_t vzip_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#if !defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vdupq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vdup_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmovq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vmovq_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) {__s0, __s0, __s0, __s0, __s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmov_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret; \
-})
-#else
-#define vmov_n_f16(__p0) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) {__s0, __s0, __s0, __s0}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#endif
-#if (__ARM_FP & 2)
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vcvt_f16_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x4_t __noswap_vcvt_f16_f32(float32x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_f32((int8x16_t)__p0, 8);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vcvt_f32_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vcvt_f32_f16(float16x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vcvt_f32_f16((int8x8_t)__p0, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f16(__p0) __extension__ ({ \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
-  __ret; \
-})
-#else
-#define vld1q_f16(__p0) __extension__ ({ \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_v(__p0, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f16(__p0) __extension__ ({ \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
-  __ret; \
-})
-#else
-#define vld1_f16(__p0) __extension__ ({ \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_v(__p0, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_f16(__p0) __extension__ ({ \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
-  __ret; \
-})
-#else
-#define vld1q_dup_f16(__p0) __extension__ ({ \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_dup_v(__p0, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_f16(__p0) __extension__ ({ \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
-  __ret; \
-})
-#else
-#define vld1_dup_f16(__p0) __extension__ ({ \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_dup_v(__p0, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
-  __ret; \
-})
-#else
-#define vld1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
-  __ret; \
-})
-#else
-#define vld1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f16_x2(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld1q_f16_x2(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f16_x2(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld1_f16_x2(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f16_x3(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld1q_f16_x3(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f16_x3(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld1_f16_x3(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f16_x4(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld1q_f16_x4(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f16_x4(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld1_f16_x4(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_f16(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld2q_f16(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_f16(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld2_f16(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_f16(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld2q_dup_f16(__p0) __extension__ ({ \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_f16(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld2_dup_f16(__p0) __extension__ ({ \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 40); \
-  __ret; \
-})
-#else
-#define vld2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  float16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 8); \
-  __ret; \
-})
-#else
-#define vld2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  float16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  float16x4x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_f16(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld3q_f16(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_f16(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld3_f16(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_f16(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld3q_dup_f16(__p0) __extension__ ({ \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_f16(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld3_dup_f16(__p0) __extension__ ({ \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 40); \
-  __ret; \
-})
-#else
-#define vld3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  float16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 8); \
-  __ret; \
-})
-#else
-#define vld3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  float16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  float16x4x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_f16(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld4q_f16(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_f16(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld4_f16(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_f16(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
-  __ret; \
-})
-#else
-#define vld4q_dup_f16(__p0) __extension__ ({ \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_f16(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
-  __ret; \
-})
-#else
-#define vld4_dup_f16(__p0) __extension__ ({ \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 40); \
-  __ret; \
-})
-#else
-#define vld4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  float16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 40); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 8); \
-  __ret; \
-})
-#else
-#define vld4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  float16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  float16x4x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 8); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 40); \
-})
-#else
-#define vst1q_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 8); \
-})
-#else
-#define vst1_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__rev1, 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 40); \
-})
-#else
-#define vst1q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 8); \
-})
-#else
-#define vst1_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__rev1, __p2, 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 40); \
-})
-#else
-#define vst1q_f16_x2(__p0, __p1) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  float16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f16_x2(__p0, __p1) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 8); \
-})
-#else
-#define vst1_f16_x2(__p0, __p1) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  float16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x2_v(__p0, __rev1.val[0], __rev1.val[1], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 40); \
-})
-#else
-#define vst1q_f16_x3(__p0, __p1) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  float16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f16_x3(__p0, __p1) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 8); \
-})
-#else
-#define vst1_f16_x3(__p0, __p1) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  float16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 40); \
-})
-#else
-#define vst1q_f16_x4(__p0, __p1) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  float16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f16_x4(__p0, __p1) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 8); \
-})
-#else
-#define vst1_f16_x4(__p0, __p1) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  float16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst1_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 40); \
-})
-#else
-#define vst2q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  float16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_f16(__p0, __p1) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 8); \
-})
-#else
-#define vst2_f16(__p0, __p1) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  float16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_v(__p0, __rev1.val[0], __rev1.val[1], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 40); \
-})
-#else
-#define vst2q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x2_t __s1 = __p1; \
-  float16x8x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 8); \
-})
-#else
-#define vst2_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x2_t __s1 = __p1; \
-  float16x4x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __builtin_neon_vst2_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 40); \
-})
-#else
-#define vst3q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  float16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_f16(__p0, __p1) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 8); \
-})
-#else
-#define vst3_f16(__p0, __p1) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  float16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 40); \
-})
-#else
-#define vst3q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x3_t __s1 = __p1; \
-  float16x8x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 8); \
-})
-#else
-#define vst3_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x3_t __s1 = __p1; \
-  float16x4x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __builtin_neon_vst3_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 40); \
-})
-#else
-#define vst4q_f16(__p0, __p1) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  float16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_f16(__p0, __p1) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 8); \
-})
-#else
-#define vst4_f16(__p0, __p1) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  float16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 8); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 40); \
-})
-#else
-#define vst4q_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8x4_t __s1 = __p1; \
-  float16x8x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 40); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 8); \
-})
-#else
-#define vst4_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4x4_t __s1 = __p1; \
-  float16x4x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 3, 2, 1, 0); \
-  __builtin_neon_vst4_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 8); \
-})
-#endif
-
-#endif
-#if __ARM_ARCH >= 8
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vcvtaq_s32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtaq_s32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcvta_s32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvta_s32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcvtaq_u32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtaq_u32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcvta_u32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvta_u32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vcvtmq_s32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtmq_s32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcvtm_s32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtm_s32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcvtmq_u32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtmq_u32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcvtm_u32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtm_u32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vcvtnq_s32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtnq_s32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcvtn_s32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtn_s32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcvtnq_u32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtnq_u32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcvtn_u32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtn_u32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__p0, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vcvtpq_s32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vcvtpq_s32_v((int8x16_t)__rev0, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__p0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vcvtp_s32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vcvtp_s32_v((int8x8_t)__rev0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcvtpq_u32_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcvtpq_u32_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcvtp_u32_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcvtp_u32_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_CRYPTO)
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesdq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vaesdq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaeseq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vaeseq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaeseq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesimcq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vaesimcq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesimcq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesmcq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vaesmcq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vaesmcq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha1cq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1cq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vsha1h_u32(uint32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vsha1h_u32(uint32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsha1h_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha1mq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1mq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32((int8x16_t)__p0, __p1, (int8x16_t)__p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha1pq_u32(uint32x4_t __p0, uint32_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1pq_u32((int8x16_t)__rev0, __p1, (int8x16_t)__rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1su0q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha1su0q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1su0q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1su1q_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha1su1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha1su1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256hq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha256hq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256hq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256h2q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha256h2q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256h2q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256su0q_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha256su0q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256su0q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256su1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsha256su1q_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsha256su1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrnd_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrnd_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrnd_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndaq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndaq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrnda_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrnda_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrnda_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndiq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndiq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrndi_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrndi_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndi_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndmq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndmq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrndm_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrndm_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndm_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndnq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndnq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrndn_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrndn_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndn_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrndns_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vrndns_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrndns_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndpq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndpq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrndp_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrndp_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndp_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vrndxq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vrndxq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vrndx_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vrndx_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vrndx_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_DIRECTED_ROUNDING) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC)
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrnd_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrnd_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrnd_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndaq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndaq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrnda_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrnda_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrnda_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndmq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndmq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrndm_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrndm_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndm_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndnq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndnq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrndn_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrndn_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndn_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndpq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndpq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrndp_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrndp_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndp_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndxq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndxq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrndx_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrndx_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndx_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vminnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__ARM_FEATURE_NUMERIC_MAXMIN) && defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC)
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vminnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vminnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vcvtaq_s64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtaq_s64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vcvta_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvta_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcvtaq_u64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtaq_u64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcvta_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvta_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vcvtmq_s64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtmq_s64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vcvtm_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtm_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcvtmq_u64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtmq_u64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcvtm_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtm_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vcvtnq_s64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtnq_s64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vcvtn_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtn_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcvtnq_u64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtnq_u64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcvtn_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtn_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vcvtpq_s64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtpq_s64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vcvtp_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvtp_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcvtpq_u64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtpq_u64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcvtp_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvtp_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_p64(poly64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_p16(poly16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u8(uint8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u32(uint32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u64(uint64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_u16(uint16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s8(int8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_f64(float64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_f32(float32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_f16(float16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s32(int32x2_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s64(int64x1_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x8_t vreinterpret_p8_s16(int16x4_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_p8(poly8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_p16(poly16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_u8(uint8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_u32(uint32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_u64(uint64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_u16(uint16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_s8(int8x8_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_f64(float64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_f32(float32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_f16(float16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_s32(int32x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_s64(int64x1_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vreinterpret_p64_s16(int16x4_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_p8(poly8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_p64(poly64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u8(uint8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u32(uint32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u64(uint64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_u16(uint16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s8(int8x8_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_f64(float64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_f32(float32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_f16(float16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s32(int32x2_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s64(int64x1_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x4_t vreinterpret_p16_s16(int16x4_t __p0) {
-  poly16x4_t __ret;
-  __ret = (poly16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_p128(poly128_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_p64(poly64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_p16(poly16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u8(uint8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u32(uint32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u64(uint64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_u16(uint16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s8(int8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_f64(float64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_f32(float32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_f16(float16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s32(int32x4_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s64(int64x2_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly8x16_t vreinterpretq_p8_s16(int16x8_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_p8(poly8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_p64(poly64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_p16(poly16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_u8(uint8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_u32(uint32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_u64(uint64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_u16(uint16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_s8(int8x16_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_f64(float64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_f32(float32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_f16(float16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_s32(int32x4_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_s64(int64x2_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly128_t vreinterpretq_p128_s16(int16x8_t __p0) {
-  poly128_t __ret;
-  __ret = (poly128_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_p8(poly8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_p128(poly128_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_p16(poly16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_u8(uint8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_u32(uint32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_u64(uint64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_u16(uint16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_s8(int8x16_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_f64(float64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_f32(float32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_f16(float16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_s32(int32x4_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_s64(int64x2_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x2_t vreinterpretq_p64_s16(int16x8_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_p8(poly8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_p128(poly128_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_p64(poly64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u8(uint8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u32(uint32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u64(uint64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_u16(uint16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s8(int8x16_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_f64(float64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_f32(float32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_f16(float16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s32(int32x4_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s64(int64x2_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly16x8_t vreinterpretq_p16_s16(int16x8_t __p0) {
-  poly16x8_t __ret;
-  __ret = (poly16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p128(poly128_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p64(poly64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_p16(poly16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u32(uint32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u64(uint64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_u16(uint16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_f64(float64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_f32(float32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_f16(float16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s32(int32x4_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s64(int64x2_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x16_t vreinterpretq_u8_s16(int16x8_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p8(poly8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p128(poly128_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p64(poly64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_p16(poly16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u8(uint8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u64(uint64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_u16(uint16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s8(int8x16_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_f64(float64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_f16(float16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s64(int64x2_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x4_t vreinterpretq_u32_s16(int16x8_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p8(poly8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p128(poly128_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p64(poly64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_p16(poly16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u8(uint8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u32(uint32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_u16(uint16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s8(int8x16_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_f32(float32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_f16(float16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s32(int32x4_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x2_t vreinterpretq_u64_s16(int16x8_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p8(poly8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p128(poly128_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p64(poly64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u8(uint8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u32(uint32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_u64(uint64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s8(int8x16_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_f64(float64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_f32(float32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s32(int32x4_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s64(int64x2_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x8_t vreinterpretq_u16_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p8(poly8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p128(poly128_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p64(poly64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_p16(poly16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u8(uint8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u32(uint32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u64(uint64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_u16(uint16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_f64(float64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_f32(float32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_f16(float16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s32(int32x4_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s64(int64x2_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x16_t vreinterpretq_s8_s16(int16x8_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_p8(poly8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_p128(poly128_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_p64(poly64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_p16(poly16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_u8(uint8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_u32(uint32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_u64(uint64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_u16(uint16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_s8(int8x16_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_f32(float32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_f16(float16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_s32(int32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_s64(int64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x2_t vreinterpretq_f64_s16(int16x8_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p8(poly8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p128(poly128_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p64(poly64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_p16(poly16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u8(uint8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u32(uint32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u64(uint64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_u16(uint16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s8(int8x16_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_f64(float64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s32(int32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s64(int64x2_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x4_t vreinterpretq_f32_s16(int16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p8(poly8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p128(poly128_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p64(poly64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_p16(poly16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u8(uint8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u32(uint32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u64(uint64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s8(int8x16_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_f64(float64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_f32(float32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s32(int32x4_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s64(int64x2_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x8_t vreinterpretq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p8(poly8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p128(poly128_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p64(poly64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_p16(poly16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u8(uint8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u32(uint32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u64(uint64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_u16(uint16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s8(int8x16_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_f64(float64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_f32(float32x4_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_f16(float16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s64(int64x2_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x4_t vreinterpretq_s32_s16(int16x8_t __p0) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p8(poly8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p128(poly128_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p64(poly64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_p16(poly16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u8(uint8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u32(uint32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u64(uint64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_u16(uint16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s8(int8x16_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_f32(float32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_f16(float16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s32(int32x4_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x2_t vreinterpretq_s64_s16(int16x8_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p8(poly8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p128(poly128_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p64(poly64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_p16(poly16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u8(uint8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u32(uint32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u64(uint64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_u16(uint16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s8(int8x16_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_f64(float64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_f32(float32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s32(int32x4_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x8_t vreinterpretq_s16_s64(int64x2_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_p64(poly64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_p16(poly16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u32(uint32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u64(uint64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_u16(uint16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_f64(float64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_f32(float32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_f16(float16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s32(int32x2_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s64(int64x1_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint8x8_t vreinterpret_u8_s16(int16x4_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_p8(poly8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_p64(poly64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_p16(poly16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u8(uint8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u64(uint64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_u16(uint16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s8(int8x8_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_f64(float64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_f16(float16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s64(int64x1_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint32x2_t vreinterpret_u32_s16(int16x4_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_p8(poly8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_p16(poly16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u8(uint8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u32(uint32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_u16(uint16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s8(int8x8_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_f32(float32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_f16(float16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s32(int32x2_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint64x1_t vreinterpret_u64_s16(int16x4_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_p8(poly8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_p64(poly64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u8(uint8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u32(uint32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_u64(uint64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s8(int8x8_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_f64(float64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_f32(float32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s32(int32x2_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s64(int64x1_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai uint16x4_t vreinterpret_u16_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_p8(poly8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_p64(poly64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_p16(poly16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u8(uint8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u32(uint32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u64(uint64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_u16(uint16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_f64(float64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_f32(float32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_f16(float16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s32(int32x2_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s64(int64x1_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#else
-__ai int8x8_t vreinterpret_s8_s16(int16x4_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_p8(poly8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_p64(poly64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_p16(poly16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_u8(uint8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_u32(uint32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_u16(uint16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_s8(int8x8_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_f32(float32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_f16(float16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_s32(int32x2_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vreinterpret_f64_s16(int16x4_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_p8(poly8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_p64(poly64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_p16(poly16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u8(uint8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u32(uint32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u64(uint64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_u16(uint16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s8(int8x8_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_f64(float64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_f16(float16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s32(int32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s64(int64x1_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai float32x2_t vreinterpret_f32_s16(int16x4_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_p8(poly8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_p64(poly64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_p16(poly16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u8(uint8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u32(uint32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u64(uint64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s8(int8x8_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_f64(float64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_f32(float32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s32(int32x2_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s64(int64x1_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai float16x4_t vreinterpret_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_p8(poly8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_p64(poly64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_p16(poly16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u8(uint8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u32(uint32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u64(uint64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_u16(uint16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s8(int8x8_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_f64(float64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_f32(float32x2_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_f16(float16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s64(int64x1_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#else
-__ai int32x2_t vreinterpret_s32_s16(int16x4_t __p0) {
-  int32x2_t __ret;
-  __ret = (int32x2_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_p8(poly8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_p64(poly64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_p16(poly16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u8(uint8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u32(uint32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u64(uint64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_u16(uint16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s8(int8x8_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_f32(float32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_f16(float16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s32(int32x2_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai int64x1_t vreinterpret_s64_s16(int16x4_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_p8(poly8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_p64(poly64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_p16(poly16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u8(uint8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u32(uint32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u64(uint64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_u16(uint16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s8(int8x8_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_f64(float64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_f32(float32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s32(int32x2_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#else
-__ai int16x4_t vreinterpret_s16_s64(int64x1_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t)(__p0);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_DIRECTED_ROUNDING)
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrnd_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrnd_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnd_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndaq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndaq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndaq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrnda_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrnda_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrnda_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndiq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndiq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrndi_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrndi_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndmq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndmq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndmq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrndm_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrndm_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndm_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndnq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndnq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndnq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrndn_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrndn_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndn_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndpq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndpq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndpq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrndp_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrndp_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndp_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrndxq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrndxq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrndxq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrndx_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrndx_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrndx_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#endif
-#if __ARM_ARCH >= 8 && defined(__aarch64__) && defined(__ARM_FEATURE_NUMERIC_MAXMIN)
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vmaxnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vminnm_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_DOTPROD)
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vdotq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vdotq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vdotq_u32(uint32x4_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vdotq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 50);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vdotq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vdotq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vdotq_s32(int32x4_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vdotq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 34);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vdot_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vdot_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint32x2_t __noswap_vdot_u32(uint32x2_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vdot_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 18);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vdot_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vdot_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int32x2_t __noswap_vdot_s32(int32x2_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vdot_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdotq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x8_t __s2 = __p2; \
-  uint32x4_t __ret; \
-uint8x8_t __reint = __s2; \
-uint32x4_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = vdotq_u32(__s0, __s1, *(uint8x16_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdotq_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x8_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-uint8x8_t __reint = __rev2; \
-uint32x4_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = __noswap_vdotq_u32(__rev0, __rev1, *(uint8x16_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdotq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-int8x8_t __reint = __s2; \
-int32x4_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = vdotq_s32(__s0, __s1, *(int8x16_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdotq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-int8x8_t __reint = __rev2; \
-int32x4_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = __noswap_vdotq_s32(__rev0, __rev1, *(int8x16_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdot_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __s2 = __p2; \
-  uint32x2_t __ret; \
-uint8x8_t __reint = __s2; \
-uint32x2_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3); \
-  __ret = vdot_u32(__s0, __s1, *(uint8x8_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdot_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x8_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-uint8x8_t __reint = __rev2; \
-uint32x2_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3); \
-  __ret = __noswap_vdot_u32(__rev0, __rev1, *(uint8x8_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdot_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __s2 = __p2; \
-  int32x2_t __ret; \
-int8x8_t __reint = __s2; \
-int32x2_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3); \
-  __ret = vdot_s32(__s0, __s1, *(int8x8_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdot_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x8_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-int8x8_t __reint = __rev2; \
-int32x2_t __reint1 = __builtin_shufflevector(*(uint32x2_t *) &__reint, *(uint32x2_t *) &__reint, __p3, __p3); \
-  __ret = __noswap_vdot_s32(__rev0, __rev1, *(int8x8_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_DOTPROD) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vdotq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __s2 = __p2; \
-  uint32x4_t __ret; \
-uint8x16_t __reint = __s2; \
-uint32x4_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = vdotq_u32(__s0, __s1, *(uint8x16_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdotq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint8x16_t __s1 = __p1; \
-  uint8x16_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-uint8x16_t __reint = __rev2; \
-uint32x4_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = __noswap_vdotq_u32(__rev0, __rev1, *(uint8x16_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdotq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __s2 = __p2; \
-  int32x4_t __ret; \
-int8x16_t __reint = __s2; \
-int32x4_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = vdotq_s32(__s0, __s1, *(int8x16_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdotq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int8x16_t __s1 = __p1; \
-  int8x16_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-int8x16_t __reint = __rev2; \
-int32x4_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3, __p3, __p3); \
-  __ret = __noswap_vdotq_s32(__rev0, __rev1, *(int8x16_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdot_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x16_t __s2 = __p2; \
-  uint32x2_t __ret; \
-uint8x16_t __reint = __s2; \
-uint32x2_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3); \
-  __ret = vdot_u32(__s0, __s1, *(uint8x8_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdot_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint8x8_t __s1 = __p1; \
-  uint8x16_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-uint8x16_t __reint = __rev2; \
-uint32x2_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3); \
-  __ret = __noswap_vdot_u32(__rev0, __rev1, *(uint8x8_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdot_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x16_t __s2 = __p2; \
-  int32x2_t __ret; \
-int8x16_t __reint = __s2; \
-int32x2_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3); \
-  __ret = vdot_s32(__s0, __s1, *(int8x8_t *) &__reint1); \
-  __ret; \
-})
-#else
-#define vdot_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int8x8_t __s1 = __p1; \
-  int8x16_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-int8x16_t __reint = __rev2; \
-int32x2_t __reint1 = __builtin_shufflevector(*(uint32x4_t *) &__reint, *(uint32x4_t *) &__reint, __p3, __p3); \
-  __ret = __noswap_vdot_s32(__rev0, __rev1, *(int8x8_t *) &__reint1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_FMA)
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vfmaq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vfma_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __ret;
-  __ret = vfmaq_f32(__p0, __p1, (float32x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai float32x4_t vfmaq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vfmaq_f32(__rev0, __rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __ret;
-  __ret = vfma_f32(__p0, __p1, (float32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai float32x2_t vfma_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __noswap_vfma_f32(__rev0, __rev1, (float32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __ret;
-  __ret = vfmaq_f32(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmsq_f32(float32x4_t __p0, float32x4_t __p1, float32x4_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vfmaq_f32(__rev0, -__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __ret;
-  __ret = vfma_f32(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float32x2_t vfms_f32(float32x2_t __p0, float32x2_t __p1, float32x2_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float32x2_t __ret;
-  __ret = __noswap_vfma_f32(__rev0, -__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_FP16FML) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_high_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_high_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vfmlalq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_high_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_high_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_high_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vfmlal_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_high_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_low_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_low_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vfmlalq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlalq_low_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_low_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_low_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vfmlal_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlal_low_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_high_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_high_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vfmlslq_high_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_high_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_high_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_high_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vfmlsl_high_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_high_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_low_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_low_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vfmlslq_low_f16(float32x4_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vfmlslq_low_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_low_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_low_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vfmlsl_low_f16(float32x2_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vfmlsl_low_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 9);
-  return __ret;
-}
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC)
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vabdq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vabdq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vabd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vabd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vabd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vabsq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vabsq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vabs_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vabs_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vabs_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vabs_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vaddq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float16x8_t vaddq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vadd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float16x4_t vadd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vbslq_f16(uint16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vbslq_f16(uint16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vbsl_f16(uint16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vbsl_f16(uint16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vbsl_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcageq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcageq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcage_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcage_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcage_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcagtq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcagtq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcagt_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcagt_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcagt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcaleq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcaleq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcale_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcale_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcale_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcaltq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcaltq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcalt_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcalt_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcalt_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceq_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceq_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqzq_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqzq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceqz_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceqz_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgeq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgeq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcge_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcge_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgezq_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgezq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgez_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgez_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgtq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgtq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgt_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgt_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgtzq_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgtzq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgtz_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgtz_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcleq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcleq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcle_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcle_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vclezq_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vclezq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclez_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclez_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcltq_f16(float16x8_t __p0, float16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcltq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclt_f16(float16x4_t __p0, float16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclt_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcltzq_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcltzq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcltz_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcltz_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vcvtq_f16_u16(uint16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vcvtq_f16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai float16x8_t vcvtq_f16_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vcvtq_f16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vcvtq_f16_s16(int16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vcvtq_f16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai float16x8_t vcvtq_f16_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vcvtq_f16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vcvt_f16_u16(uint16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai float16x4_t vcvt_f16_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vcvt_f16_s16(int16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai float16x4_t vcvt_f16_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vcvt_f16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vcvtq_n_f16_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vcvt_n_f16_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vcvt_n_f16_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vcvt_n_f16_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vcvtq_n_s16_v((int8x16_t)__s0, __p1, 33); \
-  __ret; \
-})
-#else
-#define vcvtq_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = (int16x8_t) __builtin_neon_vcvtq_n_s16_v((int8x16_t)__rev0, __p1, 33); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vcvt_n_s16_v((int8x8_t)__s0, __p1, 1); \
-  __ret; \
-})
-#else
-#define vcvt_n_s16_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = (int16x4_t) __builtin_neon_vcvt_n_s16_v((int8x8_t)__rev0, __p1, 1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vcvtq_n_u16_v((int8x16_t)__s0, __p1, 49); \
-  __ret; \
-})
-#else
-#define vcvtq_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = (uint16x8_t) __builtin_neon_vcvtq_n_u16_v((int8x16_t)__rev0, __p1, 49); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vcvt_n_u16_v((int8x8_t)__s0, __p1, 17); \
-  __ret; \
-})
-#else
-#define vcvt_n_u16_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = (uint16x4_t) __builtin_neon_vcvt_n_u16_v((int8x8_t)__rev0, __p1, 17); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcvtq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtq_s16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vcvtq_s16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtq_s16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcvt_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvt_s16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcvt_s16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvt_s16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcvtq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtq_u16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcvtq_u16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtq_u16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcvt_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvt_u16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcvt_u16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvt_u16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcvtaq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtaq_s16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vcvtaq_s16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtaq_s16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcvta_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvta_s16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcvta_s16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvta_s16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcvtaq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtaq_u16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcvtaq_u16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtaq_u16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcvta_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvta_u16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcvta_u16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvta_u16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcvtmq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtmq_s16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vcvtmq_s16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtmq_s16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcvtm_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtm_s16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcvtm_s16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtm_s16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcvtmq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtmq_u16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcvtmq_u16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtmq_u16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcvtm_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtm_u16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcvtm_u16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtm_u16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcvtnq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtnq_s16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vcvtnq_s16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtnq_s16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcvtn_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtn_s16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcvtn_s16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtn_s16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcvtnq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtnq_u16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcvtnq_u16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtnq_u16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcvtn_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtn_u16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcvtn_u16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtn_u16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vcvtpq_s16_f16(float16x8_t __p0) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtpq_s16_v((int8x16_t)__p0, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vcvtpq_s16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vcvtpq_s16_v((int8x16_t)__rev0, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vcvtp_s16_f16(float16x4_t __p0) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtp_s16_v((int8x8_t)__p0, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vcvtp_s16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vcvtp_s16_v((int8x8_t)__rev0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcvtpq_u16_f16(float16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtpq_u16_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcvtpq_u16_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcvtpq_u16_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcvtp_u16_f16(float16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtp_u16_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcvtp_u16_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcvtp_u16_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 40); \
-  __ret; \
-})
-#else
-#define vextq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 8); \
-  __ret; \
-})
-#else
-#define vext_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vext_v((int8x8_t)__rev0, (int8x8_t)__rev1, __p2, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x8_t __noswap_vfmaq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 40);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vfma_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x4_t __noswap_vfma_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 8);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vfmsq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __ret;
-  __ret = vfmaq_f16(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float16x8_t vfmsq_f16(float16x8_t __p0, float16x8_t __p1, float16x8_t __p2) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __noswap_vfmaq_f16(__rev0, -__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vfms_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __ret;
-  __ret = vfma_f16(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float16x4_t vfms_f16(float16x4_t __p0, float16x4_t __p1, float16x4_t __p2) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __noswap_vfma_f16(__rev0, -__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vmax_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vmax_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vminq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vminq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vmin_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vmin_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vmulq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float16x8_t vmulq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float16x4_t vmul_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_n_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = __s0 * (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}; \
-  __ret; \
-})
-#else
-#define vmulq_n_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __rev0 * (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_n_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = __s0 * (float16x4_t) {__s1, __s1, __s1, __s1}; \
-  __ret; \
-})
-#else
-#define vmul_n_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __rev0 * (float16x4_t) {__s1, __s1, __s1, __s1}; \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vnegq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float16x8_t vnegq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vneg_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float16x4_t vneg_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vpadd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpadd_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vpadd_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vpmax_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmax_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vpmax_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmax_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vpmin_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmin_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vpmin_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmin_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrecpeq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrecpeq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrecpe_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrecpe_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrecpe_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrecpsq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrecpsq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrecps_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrecps_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrecps_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrev64q_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0, 7, 6, 5, 4);
-  return __ret;
-}
-#else
-__ai float16x8_t vrev64q_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0, 7, 6, 5, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrev64_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  return __ret;
-}
-#else
-__ai float16x4_t vrev64_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 3, 2, 1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrsqrteq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrsqrteq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrsqrte_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrsqrte_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrsqrte_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrsqrtsq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrsqrtsq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrsqrts_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrsqrts_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrsqrts_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vsubq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float16x8_t vsubq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vsub_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float16x4_t vsub_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8x2_t vtrnq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8x2_t vtrnq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8x2_t __ret;
-  __builtin_neon_vtrnq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4x2_t vtrn_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4x2_t vtrn_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4x2_t __ret;
-  __builtin_neon_vtrn_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8x2_t vuzpq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8x2_t vuzpq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8x2_t __ret;
-  __builtin_neon_vuzpq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4x2_t vuzp_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4x2_t vuzp_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4x2_t __ret;
-  __builtin_neon_vuzp_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8x2_t vzipq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8x2_t vzipq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8x2_t __ret;
-  __builtin_neon_vzipq_v(&__ret, (int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 7, 6, 5, 4, 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4x2_t vzip_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4x2_t vzip_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4x2_t __ret;
-  __builtin_neon_vzip_v(&__ret, (int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 3, 2, 1, 0);
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vdivq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float16x8_t vdivq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __rev0 / __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float16x4_t vdiv_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __rev0 / __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vduph_lane_f16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vduph_lane_f16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vduph_laneq_f16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vduph_laneq_f16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmah_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_lane_f16(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 40); \
-  __ret; \
-})
-#else
-#define vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, __p3, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 40); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 8); \
-  __ret; \
-})
-#else
-#define vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, __p3, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfma_lane_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 8); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmah_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vfmah_laneq_f16(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 40); \
-  __ret; \
-})
-#else
-#define vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 40); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = (float16x8_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 40); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 8); \
-  __ret; \
-})
-#else
-#define vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_laneq_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x16_t)__rev2, __p3, 8); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfma_laneq_f16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = (float16x4_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 8); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = vfmaq_f16(__s0, __s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
-  __ret; \
-})
-#else
-#define vfmaq_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __noswap_vfmaq_f16(__rev0, __rev1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = vfma_f16(__s0, __s1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
-  __ret; \
-})
-#else
-#define vfma_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __noswap_vfma_f16(__rev0, __rev1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsh_lane_f16(__p0_0, __p1_0, __p2_0, __p3_0) __extension__ ({ \
-  float16_t __s0_0 = __p0_0; \
-  float16_t __s1_0 = __p1_0; \
-  float16x4_t __s2_0 = __p2_0; \
-  float16_t __ret_0; \
-  __ret_0 = vfmah_lane_f16(__s0_0, -__s1_0, __s2_0, __p3_0); \
-  __ret_0; \
-})
-#else
-#define vfmsh_lane_f16(__p0_1, __p1_1, __p2_1, __p3_1) __extension__ ({ \
-  float16_t __s0_1 = __p0_1; \
-  float16_t __s1_1 = __p1_1; \
-  float16x4_t __s2_1 = __p2_1; \
-  float16x4_t __rev2_1;  __rev2_1 = __builtin_shufflevector(__s2_1, __s2_1, 3, 2, 1, 0); \
-  float16_t __ret_1; \
-  __ret_1 = __noswap_vfmah_lane_f16(__s0_1, -__s1_1, __rev2_1, __p3_1); \
-  __ret_1; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f16(__p0_2, __p1_2, __p2_2, __p3_2) __extension__ ({ \
-  float16x8_t __s0_2 = __p0_2; \
-  float16x8_t __s1_2 = __p1_2; \
-  float16x4_t __s2_2 = __p2_2; \
-  float16x8_t __ret_2; \
-  __ret_2 = vfmaq_lane_f16(__s0_2, -__s1_2, __s2_2, __p3_2); \
-  __ret_2; \
-})
-#else
-#define vfmsq_lane_f16(__p0_3, __p1_3, __p2_3, __p3_3) __extension__ ({ \
-  float16x8_t __s0_3 = __p0_3; \
-  float16x8_t __s1_3 = __p1_3; \
-  float16x4_t __s2_3 = __p2_3; \
-  float16x8_t __rev0_3;  __rev0_3 = __builtin_shufflevector(__s0_3, __s0_3, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1_3;  __rev1_3 = __builtin_shufflevector(__s1_3, __s1_3, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_3;  __rev2_3 = __builtin_shufflevector(__s2_3, __s2_3, 3, 2, 1, 0); \
-  float16x8_t __ret_3; \
-  __ret_3 = __noswap_vfmaq_lane_f16(__rev0_3, -__rev1_3, __rev2_3, __p3_3); \
-  __ret_3 = __builtin_shufflevector(__ret_3, __ret_3, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_3; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f16(__p0_4, __p1_4, __p2_4, __p3_4) __extension__ ({ \
-  float16x4_t __s0_4 = __p0_4; \
-  float16x4_t __s1_4 = __p1_4; \
-  float16x4_t __s2_4 = __p2_4; \
-  float16x4_t __ret_4; \
-  __ret_4 = vfma_lane_f16(__s0_4, -__s1_4, __s2_4, __p3_4); \
-  __ret_4; \
-})
-#else
-#define vfms_lane_f16(__p0_5, __p1_5, __p2_5, __p3_5) __extension__ ({ \
-  float16x4_t __s0_5 = __p0_5; \
-  float16x4_t __s1_5 = __p1_5; \
-  float16x4_t __s2_5 = __p2_5; \
-  float16x4_t __rev0_5;  __rev0_5 = __builtin_shufflevector(__s0_5, __s0_5, 3, 2, 1, 0); \
-  float16x4_t __rev1_5;  __rev1_5 = __builtin_shufflevector(__s1_5, __s1_5, 3, 2, 1, 0); \
-  float16x4_t __rev2_5;  __rev2_5 = __builtin_shufflevector(__s2_5, __s2_5, 3, 2, 1, 0); \
-  float16x4_t __ret_5; \
-  __ret_5 = __noswap_vfma_lane_f16(__rev0_5, -__rev1_5, __rev2_5, __p3_5); \
-  __ret_5 = __builtin_shufflevector(__ret_5, __ret_5, 3, 2, 1, 0); \
-  __ret_5; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsh_laneq_f16(__p0_6, __p1_6, __p2_6, __p3_6) __extension__ ({ \
-  float16_t __s0_6 = __p0_6; \
-  float16_t __s1_6 = __p1_6; \
-  float16x8_t __s2_6 = __p2_6; \
-  float16_t __ret_6; \
-  __ret_6 = vfmah_laneq_f16(__s0_6, -__s1_6, __s2_6, __p3_6); \
-  __ret_6; \
-})
-#else
-#define vfmsh_laneq_f16(__p0_7, __p1_7, __p2_7, __p3_7) __extension__ ({ \
-  float16_t __s0_7 = __p0_7; \
-  float16_t __s1_7 = __p1_7; \
-  float16x8_t __s2_7 = __p2_7; \
-  float16x8_t __rev2_7;  __rev2_7 = __builtin_shufflevector(__s2_7, __s2_7, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_7; \
-  __ret_7 = __noswap_vfmah_laneq_f16(__s0_7, -__s1_7, __rev2_7, __p3_7); \
-  __ret_7; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f16(__p0_8, __p1_8, __p2_8, __p3_8) __extension__ ({ \
-  float16x8_t __s0_8 = __p0_8; \
-  float16x8_t __s1_8 = __p1_8; \
-  float16x8_t __s2_8 = __p2_8; \
-  float16x8_t __ret_8; \
-  __ret_8 = vfmaq_laneq_f16(__s0_8, -__s1_8, __s2_8, __p3_8); \
-  __ret_8; \
-})
-#else
-#define vfmsq_laneq_f16(__p0_9, __p1_9, __p2_9, __p3_9) __extension__ ({ \
-  float16x8_t __s0_9 = __p0_9; \
-  float16x8_t __s1_9 = __p1_9; \
-  float16x8_t __s2_9 = __p2_9; \
-  float16x8_t __rev0_9;  __rev0_9 = __builtin_shufflevector(__s0_9, __s0_9, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1_9;  __rev1_9 = __builtin_shufflevector(__s1_9, __s1_9, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_9;  __rev2_9 = __builtin_shufflevector(__s2_9, __s2_9, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret_9; \
-  __ret_9 = __noswap_vfmaq_laneq_f16(__rev0_9, -__rev1_9, __rev2_9, __p3_9); \
-  __ret_9 = __builtin_shufflevector(__ret_9, __ret_9, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_9; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f16(__p0_10, __p1_10, __p2_10, __p3_10) __extension__ ({ \
-  float16x4_t __s0_10 = __p0_10; \
-  float16x4_t __s1_10 = __p1_10; \
-  float16x8_t __s2_10 = __p2_10; \
-  float16x4_t __ret_10; \
-  __ret_10 = vfma_laneq_f16(__s0_10, -__s1_10, __s2_10, __p3_10); \
-  __ret_10; \
-})
-#else
-#define vfms_laneq_f16(__p0_11, __p1_11, __p2_11, __p3_11) __extension__ ({ \
-  float16x4_t __s0_11 = __p0_11; \
-  float16x4_t __s1_11 = __p1_11; \
-  float16x8_t __s2_11 = __p2_11; \
-  float16x4_t __rev0_11;  __rev0_11 = __builtin_shufflevector(__s0_11, __s0_11, 3, 2, 1, 0); \
-  float16x4_t __rev1_11;  __rev1_11 = __builtin_shufflevector(__s1_11, __s1_11, 3, 2, 1, 0); \
-  float16x8_t __rev2_11;  __rev2_11 = __builtin_shufflevector(__s2_11, __s2_11, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __ret_11; \
-  __ret_11 = __noswap_vfma_laneq_f16(__rev0_11, -__rev1_11, __rev2_11, __p3_11); \
-  __ret_11 = __builtin_shufflevector(__ret_11, __ret_11, 3, 2, 1, 0); \
-  __ret_11; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x8_t __ret; \
-  __ret = vfmaq_f16(__s0, -__s1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
-  __ret; \
-})
-#else
-#define vfmsq_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __noswap_vfmaq_f16(__rev0, -__rev1, (float16x8_t) {__s2, __s2, __s2, __s2, __s2, __s2, __s2, __s2}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x4_t __ret; \
-  __ret = vfma_f16(__s0, -__s1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
-  __ret; \
-})
-#else
-#define vfms_n_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16_t __s2 = __p2; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __noswap_vfma_f16(__rev0, -__rev1, (float16x4_t) {__s2, __s2, __s2, __s2}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxnmvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmvq_f16((int8x16_t)__s0); \
-  __ret; \
-})
-#else
-#define vmaxnmvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmvq_f16((int8x16_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxnmv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmv_f16((int8x8_t)__s0); \
-  __ret; \
-})
-#else
-#define vmaxnmv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxnmv_f16((int8x8_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxvq_f16((int8x16_t)__s0); \
-  __ret; \
-})
-#else
-#define vmaxvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxvq_f16((int8x16_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmaxv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxv_f16((int8x8_t)__s0); \
-  __ret; \
-})
-#else
-#define vmaxv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmaxv_f16((int8x8_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminnmvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmvq_f16((int8x16_t)__s0); \
-  __ret; \
-})
-#else
-#define vminnmvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmvq_f16((int8x16_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminnmv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmv_f16((int8x8_t)__s0); \
-  __ret; \
-})
-#else
-#define vminnmv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminnmv_f16((int8x8_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminvq_f16((int8x16_t)__s0); \
-  __ret; \
-})
-#else
-#define vminvq_f16(__p0) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminvq_f16((int8x16_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vminv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminv_f16((int8x8_t)__s0); \
-  __ret; \
-})
-#else
-#define vminv_f16(__p0) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vminv_f16((int8x8_t)__rev0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x8_t __noswap_vmulxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmulx_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float16x4_t __noswap_vmulx_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxh_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_lane_f16(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vmulxh_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_lane_f16(__s0, (int8x8_t)__rev1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = vmulxq_f16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __noswap_vmulxq_f16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = vmulx_f16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulx_lane_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __noswap_vmulx_f16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxh_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_laneq_f16(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vmulxh_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret; \
-  __ret = (float16_t) __builtin_neon_vmulxh_laneq_f16(__s0, (int8x16_t)__rev1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = vmulxq_f16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __noswap_vmulxq_f16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = vmulx_f16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulx_laneq_f16(__p0, __p1, __p2) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __noswap_vmulx_f16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_n_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __ret; \
-  __ret = vmulxq_f16(__s0, (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}); \
-  __ret; \
-})
-#else
-#define vmulxq_n_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __noswap_vmulxq_f16(__rev0, (float16x8_t) {__s1, __s1, __s1, __s1, __s1, __s1, __s1, __s1}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_n_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __ret; \
-  __ret = vmulx_f16(__s0, (float16x4_t) {__s1, __s1, __s1, __s1}); \
-  __ret; \
-})
-#else
-#define vmulx_n_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16_t __s1 = __p1; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __noswap_vmulx_f16(__rev0, (float16x4_t) {__s1, __s1, __s1, __s1}); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vpaddq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vpaddq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vpmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vpmaxq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vpmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vpmaxnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vpmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vpmaxnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vpminq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vpminq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vpminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vpminnmq_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vpminnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vpminnm_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vpminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vrndiq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndiq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vrndiq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vrndiq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vrndi_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndi_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vrndi_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vrndi_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vsqrtq_f16(float16x8_t __p0) {
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 40);
-  return __ret;
-}
-#else
-__ai float16x8_t vsqrtq_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = (float16x8_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 40);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vsqrt_f16(float16x4_t __p0) {
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 8);
-  return __ret;
-}
-#else
-__ai float16x4_t vsqrt_f16(float16x4_t __p0) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = (float16x4_t) __builtin_neon_vsqrt_v((int8x8_t)__rev0, 8);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vtrn1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai float16x8_t vtrn1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vtrn1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai float16x4_t vtrn1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vtrn2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai float16x8_t vtrn2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vtrn2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai float16x4_t vtrn2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vuzp1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai float16x8_t vuzp1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vuzp1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai float16x4_t vuzp1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vuzp2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai float16x8_t vuzp2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vuzp2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai float16x4_t vuzp2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vzip1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai float16x8_t vzip1q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vzip1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai float16x4_t vzip1_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vzip2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai float16x8_t vzip2q_f16(float16x8_t __p0, float16x8_t __p1) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x4_t vzip2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai float16x4_t vzip2_f16(float16x4_t __p0, float16x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_QRDMX)
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = vqaddq_s32(__p0, vqrdmulhq_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vqrdmlahq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqaddq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = vqaddq_s16(__p0, vqrdmulhq_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vqrdmlahq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vqaddq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = vqadd_s32(__p0, vqrdmulh_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x2_t vqrdmlah_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = __noswap_vqadd_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = vqadd_s16(__p0, vqrdmulh_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x4_t vqrdmlah_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __noswap_vqadd_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqaddq_s32(__s0, vqrdmulhq_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlahq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqaddq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = vqaddq_s16(__s0, vqrdmulhq_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlahq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqaddq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlah_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = vqadd_s32(__s0, vqrdmulh_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlah_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqadd_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlah_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = vqadd_s16(__s0, vqrdmulh_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlah_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqadd_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = vqsubq_s32(__p0, vqrdmulhq_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vqrdmlshq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqsubq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = vqsubq_s16(__p0, vqrdmulhq_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vqrdmlshq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vqsubq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = vqsub_s32(__p0, vqrdmulh_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x2_t vqrdmlsh_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = __noswap_vqsub_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = vqsub_s16(__p0, vqrdmulh_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x4_t vqrdmlsh_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __noswap_vqsub_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqsubq_s32(__s0, vqrdmulhq_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlshq_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqsubq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = vqsubq_s16(__s0, vqrdmulhq_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlshq_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqsubq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlsh_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = vqsub_s32(__s0, vqrdmulh_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlsh_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqsub_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlsh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = vqsub_s16(__s0, vqrdmulh_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlsh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqsub_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqaddq_s32(__s0, vqrdmulhq_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlahq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqaddq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = vqaddq_s16(__s0, vqrdmulhq_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlahq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqaddq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlah_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = vqadd_s32(__s0, vqrdmulh_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlah_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqadd_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlah_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = vqadd_s16(__s0, vqrdmulh_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlah_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqadd_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqsubq_s32(__s0, vqrdmulhq_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlshq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqsubq_s32(__rev0, __noswap_vqrdmulhq_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = vqsubq_s16(__s0, vqrdmulhq_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlshq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqsubq_s16(__rev0, __noswap_vqrdmulhq_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlsh_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = vqsub_s32(__s0, vqrdmulh_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlsh_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqsub_s32(__rev0, __noswap_vqrdmulh_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlsh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = vqsub_s16(__s0, vqrdmulh_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3))); \
-  __ret; \
-})
-#else
-#define vqrdmlsh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqsub_s16(__rev0, __noswap_vqrdmulh_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3))); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#endif
-#if defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vabdq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vabdq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vabd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabd_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float64_t vabdd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vabdd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float32_t vabds_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vabds_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vabsq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vabsq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vabsq_s64(int64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vabsq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vabsq_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vabs_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vabs_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float64x2_t vaddq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 + __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#else
-__ai float64x1_t vadd_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vaddhn_u32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vaddhn_u32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vaddhn_u64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vaddhn_u64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vaddhn_u16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint8x16_t vaddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vaddhn_u16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vaddhn_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vaddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vaddhn_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vaddhn_s64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vaddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vaddhn_s64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vaddhn_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int8x16_t vaddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vaddhn_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddlvq_u8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vaddlvq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddlvq_u8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddlvq_u32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vaddlvq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddlvq_u32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddlvq_u16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vaddlvq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddlvq_u16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vaddlvq_s8(int8x16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddlvq_s8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vaddlvq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddlvq_s8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vaddlvq_s32(int32x4_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddlvq_s32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int64_t vaddlvq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddlvq_s32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vaddlvq_s16(int16x8_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddlvq_s16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vaddlvq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddlvq_s16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vaddlv_u8(uint8x8_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddlv_u8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vaddlv_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddlv_u8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vaddlv_u32(uint32x2_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddlv_u32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vaddlv_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddlv_u32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vaddlv_u16(uint16x4_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddlv_u16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vaddlv_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddlv_u16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vaddlv_s8(int8x8_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddlv_s8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vaddlv_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddlv_s8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vaddlv_s32(int32x2_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddlv_s32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int64_t vaddlv_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddlv_s32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vaddlv_s16(int16x4_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddlv_s16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vaddlv_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddlv_s16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vaddvq_u8(uint8x16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vaddvq_u8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vaddvq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vaddvq_u8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vaddvq_u32(uint32x4_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddvq_u32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vaddvq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddvq_u32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vaddvq_u64(uint64x2_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddvq_u64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vaddvq_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vaddvq_u64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vaddvq_u16(uint16x8_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddvq_u16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vaddvq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddvq_u16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vaddvq_s8(int8x16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vaddvq_s8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vaddvq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vaddvq_s8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vaddvq_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vaddvq_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vaddvq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vaddvq_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vaddvq_f32(float32x4_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vaddvq_f32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vaddvq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vaddvq_f32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vaddvq_s32(int32x4_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddvq_s32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vaddvq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddvq_s32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vaddvq_s64(int64x2_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddvq_s64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int64_t vaddvq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vaddvq_s64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vaddvq_s16(int16x8_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddvq_s16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vaddvq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddvq_s16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vaddv_u8(uint8x8_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vaddv_u8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vaddv_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vaddv_u8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vaddv_u32(uint32x2_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddv_u32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vaddv_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vaddv_u32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vaddv_u16(uint16x4_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddv_u16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vaddv_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vaddv_u16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vaddv_s8(int8x8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vaddv_s8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vaddv_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vaddv_s8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vaddv_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vaddv_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vaddv_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vaddv_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vaddv_s32(int32x2_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddv_s32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vaddv_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vaddv_s32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vaddv_s16(int16x4_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddv_s16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vaddv_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vaddv_s16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
-  return __ret;
-}
-#else
-__ai poly64x1_t vbsl_p64(uint64x1_t __p0, poly64x1_t __p1, poly64x1_t __p2) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 6);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 38);
-  return __ret;
-}
-#else
-__ai poly64x2_t vbslq_p64(uint64x2_t __p0, poly64x2_t __p1, poly64x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 38);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vbslq_f64(uint64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vbslq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vbsl_f64(uint64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vbsl_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcageq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcageq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcage_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcage_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcaged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcages_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcages_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcagtq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcagtq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcagt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcagt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcagtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcagtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcagts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcagts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcaleq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcaleq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcale_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcale_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcaled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcales_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcales_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcaltq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcaltq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcalt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcalt_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcaltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcaltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcalts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcalts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceq_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqq_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 == __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceq_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceq_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceq_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 == __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vceqd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vceqd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vceqd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vceqs_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceqz_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceqz_p64(poly64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceqz_p16(poly16x4_t __p0) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqzq_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqzq_p64(poly64x2_t __p0) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqzq_p16(poly16x8_t __p0) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqzq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqzq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqzq_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqzq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vceqzq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqzq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqzq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vceqzq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vceqzq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vceqzq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vceqzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceqz_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceqz_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceqz_u64(uint64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceqz_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vceqz_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vceqz_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceqz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceqz_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceqz_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vceqz_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vceqz_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vceqz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vceqz_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vceqz_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vceqz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vceqzd_u64(uint64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vceqzd_u64(uint64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vceqzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vceqzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vceqzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vceqzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vceqzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vceqzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vceqzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vceqzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vceqzs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgeq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgeq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgeq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 >= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcge_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcge_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcge_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 >= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vcged_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcged_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcged_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcged_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcged_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcges_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcges_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgezq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgezq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgezq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgezq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgezq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgezq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgezq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcgez_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcgez_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgez_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgez_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgez_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgez_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgez_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgez_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgez_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcgezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcgezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcgezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcgezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcgezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcgezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgezs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgtq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgtq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgtq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 > __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 > __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vcgtd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcgtd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcgtd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcgts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcgtzq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgtzq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgtzq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcgtzq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcgtzq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcgtzq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcgtzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcgtz_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgtz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgtz_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcgtz_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcgtz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcgtz_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcgtz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcgtzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcgtzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcgtzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcgtzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcgtzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcgtzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcgtzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcgtzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcgtzs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcleq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcleq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcleq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 <= __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcle_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcle_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcle_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 <= __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcled_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vcled_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcled_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcled_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcled_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vcles_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcles_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vclezq_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vclezq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vclezq_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vclezq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vclezq_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vclezq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vclezq_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vclezq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vclezq_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vclezq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vclezq_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vclezq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vclezq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vclez_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vclez_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vclez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vclez_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclez_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclez_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vclez_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vclez_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vclez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vclez_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vclez_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vclez_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vclez_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vclez_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vclezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vclezd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vclezd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vclezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vclezd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vclezd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vclezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vclezs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclezs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcltq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcltq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcltq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__rev0 < __rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vclt_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vclt_f64(float64x1_t __p0, float64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#else
-__ai uint64x1_t vclt_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t)(__p0 < __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcltd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vcltd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vcltd_f64(float64_t __p0, float64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vclts_f32(float32_t __p0, float32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vclts_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vcltzq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcltzq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcltzq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vcltzq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcltzq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__p0, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vcltzq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vcltzq_v((int8x16_t)__rev0, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vcltz_s8(int8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vcltz_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcltz_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcltz_f32(float32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcltz_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vcltz_s32(int32x2_t __p0) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vcltz_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcltz_s64(int64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vcltz_s16(int16x4_t __p0) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__p0, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vcltz_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vcltz_v((int8x8_t)__rev0, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcltzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcltzd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcltzd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcltzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcltzd_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcltzd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcltzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcltzs_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcltzs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  return __ret;
-}
-#else
-__ai poly64x2_t vcombine_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  return __ret;
-}
-#else
-__ai float64x2_t vcombine_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_p8(__p0_12, __p1_12, __p2_12, __p3_12) __extension__ ({ \
-  poly8x16_t __s0_12 = __p0_12; \
-  poly8x8_t __s2_12 = __p2_12; \
-  poly8x16_t __ret_12; \
-  __ret_12 = vsetq_lane_p8(vget_lane_p8(__s2_12, __p3_12), __s0_12, __p1_12); \
-  __ret_12; \
-})
-#else
-#define vcopyq_lane_p8(__p0_13, __p1_13, __p2_13, __p3_13) __extension__ ({ \
-  poly8x16_t __s0_13 = __p0_13; \
-  poly8x8_t __s2_13 = __p2_13; \
-  poly8x16_t __rev0_13;  __rev0_13 = __builtin_shufflevector(__s0_13, __s0_13, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __rev2_13;  __rev2_13 = __builtin_shufflevector(__s2_13, __s2_13, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret_13; \
-  __ret_13 = __noswap_vsetq_lane_p8(__noswap_vget_lane_p8(__rev2_13, __p3_13), __rev0_13, __p1_13); \
-  __ret_13 = __builtin_shufflevector(__ret_13, __ret_13, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_13; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_p16(__p0_14, __p1_14, __p2_14, __p3_14) __extension__ ({ \
-  poly16x8_t __s0_14 = __p0_14; \
-  poly16x4_t __s2_14 = __p2_14; \
-  poly16x8_t __ret_14; \
-  __ret_14 = vsetq_lane_p16(vget_lane_p16(__s2_14, __p3_14), __s0_14, __p1_14); \
-  __ret_14; \
-})
-#else
-#define vcopyq_lane_p16(__p0_15, __p1_15, __p2_15, __p3_15) __extension__ ({ \
-  poly16x8_t __s0_15 = __p0_15; \
-  poly16x4_t __s2_15 = __p2_15; \
-  poly16x8_t __rev0_15;  __rev0_15 = __builtin_shufflevector(__s0_15, __s0_15, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x4_t __rev2_15;  __rev2_15 = __builtin_shufflevector(__s2_15, __s2_15, 3, 2, 1, 0); \
-  poly16x8_t __ret_15; \
-  __ret_15 = __noswap_vsetq_lane_p16(__noswap_vget_lane_p16(__rev2_15, __p3_15), __rev0_15, __p1_15); \
-  __ret_15 = __builtin_shufflevector(__ret_15, __ret_15, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_15; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_u8(__p0_16, __p1_16, __p2_16, __p3_16) __extension__ ({ \
-  uint8x16_t __s0_16 = __p0_16; \
-  uint8x8_t __s2_16 = __p2_16; \
-  uint8x16_t __ret_16; \
-  __ret_16 = vsetq_lane_u8(vget_lane_u8(__s2_16, __p3_16), __s0_16, __p1_16); \
-  __ret_16; \
-})
-#else
-#define vcopyq_lane_u8(__p0_17, __p1_17, __p2_17, __p3_17) __extension__ ({ \
-  uint8x16_t __s0_17 = __p0_17; \
-  uint8x8_t __s2_17 = __p2_17; \
-  uint8x16_t __rev0_17;  __rev0_17 = __builtin_shufflevector(__s0_17, __s0_17, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev2_17;  __rev2_17 = __builtin_shufflevector(__s2_17, __s2_17, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_17; \
-  __ret_17 = __noswap_vsetq_lane_u8(__noswap_vget_lane_u8(__rev2_17, __p3_17), __rev0_17, __p1_17); \
-  __ret_17 = __builtin_shufflevector(__ret_17, __ret_17, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_17; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_u32(__p0_18, __p1_18, __p2_18, __p3_18) __extension__ ({ \
-  uint32x4_t __s0_18 = __p0_18; \
-  uint32x2_t __s2_18 = __p2_18; \
-  uint32x4_t __ret_18; \
-  __ret_18 = vsetq_lane_u32(vget_lane_u32(__s2_18, __p3_18), __s0_18, __p1_18); \
-  __ret_18; \
-})
-#else
-#define vcopyq_lane_u32(__p0_19, __p1_19, __p2_19, __p3_19) __extension__ ({ \
-  uint32x4_t __s0_19 = __p0_19; \
-  uint32x2_t __s2_19 = __p2_19; \
-  uint32x4_t __rev0_19;  __rev0_19 = __builtin_shufflevector(__s0_19, __s0_19, 3, 2, 1, 0); \
-  uint32x2_t __rev2_19;  __rev2_19 = __builtin_shufflevector(__s2_19, __s2_19, 1, 0); \
-  uint32x4_t __ret_19; \
-  __ret_19 = __noswap_vsetq_lane_u32(__noswap_vget_lane_u32(__rev2_19, __p3_19), __rev0_19, __p1_19); \
-  __ret_19 = __builtin_shufflevector(__ret_19, __ret_19, 3, 2, 1, 0); \
-  __ret_19; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_u64(__p0_20, __p1_20, __p2_20, __p3_20) __extension__ ({ \
-  uint64x2_t __s0_20 = __p0_20; \
-  uint64x1_t __s2_20 = __p2_20; \
-  uint64x2_t __ret_20; \
-  __ret_20 = vsetq_lane_u64(vget_lane_u64(__s2_20, __p3_20), __s0_20, __p1_20); \
-  __ret_20; \
-})
-#else
-#define vcopyq_lane_u64(__p0_21, __p1_21, __p2_21, __p3_21) __extension__ ({ \
-  uint64x2_t __s0_21 = __p0_21; \
-  uint64x1_t __s2_21 = __p2_21; \
-  uint64x2_t __rev0_21;  __rev0_21 = __builtin_shufflevector(__s0_21, __s0_21, 1, 0); \
-  uint64x2_t __ret_21; \
-  __ret_21 = __noswap_vsetq_lane_u64(__noswap_vget_lane_u64(__s2_21, __p3_21), __rev0_21, __p1_21); \
-  __ret_21 = __builtin_shufflevector(__ret_21, __ret_21, 1, 0); \
-  __ret_21; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_u16(__p0_22, __p1_22, __p2_22, __p3_22) __extension__ ({ \
-  uint16x8_t __s0_22 = __p0_22; \
-  uint16x4_t __s2_22 = __p2_22; \
-  uint16x8_t __ret_22; \
-  __ret_22 = vsetq_lane_u16(vget_lane_u16(__s2_22, __p3_22), __s0_22, __p1_22); \
-  __ret_22; \
-})
-#else
-#define vcopyq_lane_u16(__p0_23, __p1_23, __p2_23, __p3_23) __extension__ ({ \
-  uint16x8_t __s0_23 = __p0_23; \
-  uint16x4_t __s2_23 = __p2_23; \
-  uint16x8_t __rev0_23;  __rev0_23 = __builtin_shufflevector(__s0_23, __s0_23, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev2_23;  __rev2_23 = __builtin_shufflevector(__s2_23, __s2_23, 3, 2, 1, 0); \
-  uint16x8_t __ret_23; \
-  __ret_23 = __noswap_vsetq_lane_u16(__noswap_vget_lane_u16(__rev2_23, __p3_23), __rev0_23, __p1_23); \
-  __ret_23 = __builtin_shufflevector(__ret_23, __ret_23, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_23; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_s8(__p0_24, __p1_24, __p2_24, __p3_24) __extension__ ({ \
-  int8x16_t __s0_24 = __p0_24; \
-  int8x8_t __s2_24 = __p2_24; \
-  int8x16_t __ret_24; \
-  __ret_24 = vsetq_lane_s8(vget_lane_s8(__s2_24, __p3_24), __s0_24, __p1_24); \
-  __ret_24; \
-})
-#else
-#define vcopyq_lane_s8(__p0_25, __p1_25, __p2_25, __p3_25) __extension__ ({ \
-  int8x16_t __s0_25 = __p0_25; \
-  int8x8_t __s2_25 = __p2_25; \
-  int8x16_t __rev0_25;  __rev0_25 = __builtin_shufflevector(__s0_25, __s0_25, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2_25;  __rev2_25 = __builtin_shufflevector(__s2_25, __s2_25, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_25; \
-  __ret_25 = __noswap_vsetq_lane_s8(__noswap_vget_lane_s8(__rev2_25, __p3_25), __rev0_25, __p1_25); \
-  __ret_25 = __builtin_shufflevector(__ret_25, __ret_25, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_25; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_f32(__p0_26, __p1_26, __p2_26, __p3_26) __extension__ ({ \
-  float32x4_t __s0_26 = __p0_26; \
-  float32x2_t __s2_26 = __p2_26; \
-  float32x4_t __ret_26; \
-  __ret_26 = vsetq_lane_f32(vget_lane_f32(__s2_26, __p3_26), __s0_26, __p1_26); \
-  __ret_26; \
-})
-#else
-#define vcopyq_lane_f32(__p0_27, __p1_27, __p2_27, __p3_27) __extension__ ({ \
-  float32x4_t __s0_27 = __p0_27; \
-  float32x2_t __s2_27 = __p2_27; \
-  float32x4_t __rev0_27;  __rev0_27 = __builtin_shufflevector(__s0_27, __s0_27, 3, 2, 1, 0); \
-  float32x2_t __rev2_27;  __rev2_27 = __builtin_shufflevector(__s2_27, __s2_27, 1, 0); \
-  float32x4_t __ret_27; \
-  __ret_27 = __noswap_vsetq_lane_f32(__noswap_vget_lane_f32(__rev2_27, __p3_27), __rev0_27, __p1_27); \
-  __ret_27 = __builtin_shufflevector(__ret_27, __ret_27, 3, 2, 1, 0); \
-  __ret_27; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_s32(__p0_28, __p1_28, __p2_28, __p3_28) __extension__ ({ \
-  int32x4_t __s0_28 = __p0_28; \
-  int32x2_t __s2_28 = __p2_28; \
-  int32x4_t __ret_28; \
-  __ret_28 = vsetq_lane_s32(vget_lane_s32(__s2_28, __p3_28), __s0_28, __p1_28); \
-  __ret_28; \
-})
-#else
-#define vcopyq_lane_s32(__p0_29, __p1_29, __p2_29, __p3_29) __extension__ ({ \
-  int32x4_t __s0_29 = __p0_29; \
-  int32x2_t __s2_29 = __p2_29; \
-  int32x4_t __rev0_29;  __rev0_29 = __builtin_shufflevector(__s0_29, __s0_29, 3, 2, 1, 0); \
-  int32x2_t __rev2_29;  __rev2_29 = __builtin_shufflevector(__s2_29, __s2_29, 1, 0); \
-  int32x4_t __ret_29; \
-  __ret_29 = __noswap_vsetq_lane_s32(__noswap_vget_lane_s32(__rev2_29, __p3_29), __rev0_29, __p1_29); \
-  __ret_29 = __builtin_shufflevector(__ret_29, __ret_29, 3, 2, 1, 0); \
-  __ret_29; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_s64(__p0_30, __p1_30, __p2_30, __p3_30) __extension__ ({ \
-  int64x2_t __s0_30 = __p0_30; \
-  int64x1_t __s2_30 = __p2_30; \
-  int64x2_t __ret_30; \
-  __ret_30 = vsetq_lane_s64(vget_lane_s64(__s2_30, __p3_30), __s0_30, __p1_30); \
-  __ret_30; \
-})
-#else
-#define vcopyq_lane_s64(__p0_31, __p1_31, __p2_31, __p3_31) __extension__ ({ \
-  int64x2_t __s0_31 = __p0_31; \
-  int64x1_t __s2_31 = __p2_31; \
-  int64x2_t __rev0_31;  __rev0_31 = __builtin_shufflevector(__s0_31, __s0_31, 1, 0); \
-  int64x2_t __ret_31; \
-  __ret_31 = __noswap_vsetq_lane_s64(__noswap_vget_lane_s64(__s2_31, __p3_31), __rev0_31, __p1_31); \
-  __ret_31 = __builtin_shufflevector(__ret_31, __ret_31, 1, 0); \
-  __ret_31; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_s16(__p0_32, __p1_32, __p2_32, __p3_32) __extension__ ({ \
-  int16x8_t __s0_32 = __p0_32; \
-  int16x4_t __s2_32 = __p2_32; \
-  int16x8_t __ret_32; \
-  __ret_32 = vsetq_lane_s16(vget_lane_s16(__s2_32, __p3_32), __s0_32, __p1_32); \
-  __ret_32; \
-})
-#else
-#define vcopyq_lane_s16(__p0_33, __p1_33, __p2_33, __p3_33) __extension__ ({ \
-  int16x8_t __s0_33 = __p0_33; \
-  int16x4_t __s2_33 = __p2_33; \
-  int16x8_t __rev0_33;  __rev0_33 = __builtin_shufflevector(__s0_33, __s0_33, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2_33;  __rev2_33 = __builtin_shufflevector(__s2_33, __s2_33, 3, 2, 1, 0); \
-  int16x8_t __ret_33; \
-  __ret_33 = __noswap_vsetq_lane_s16(__noswap_vget_lane_s16(__rev2_33, __p3_33), __rev0_33, __p1_33); \
-  __ret_33 = __builtin_shufflevector(__ret_33, __ret_33, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_33; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_p8(__p0_34, __p1_34, __p2_34, __p3_34) __extension__ ({ \
-  poly8x8_t __s0_34 = __p0_34; \
-  poly8x8_t __s2_34 = __p2_34; \
-  poly8x8_t __ret_34; \
-  __ret_34 = vset_lane_p8(vget_lane_p8(__s2_34, __p3_34), __s0_34, __p1_34); \
-  __ret_34; \
-})
-#else
-#define vcopy_lane_p8(__p0_35, __p1_35, __p2_35, __p3_35) __extension__ ({ \
-  poly8x8_t __s0_35 = __p0_35; \
-  poly8x8_t __s2_35 = __p2_35; \
-  poly8x8_t __rev0_35;  __rev0_35 = __builtin_shufflevector(__s0_35, __s0_35, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __rev2_35;  __rev2_35 = __builtin_shufflevector(__s2_35, __s2_35, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret_35; \
-  __ret_35 = __noswap_vset_lane_p8(__noswap_vget_lane_p8(__rev2_35, __p3_35), __rev0_35, __p1_35); \
-  __ret_35 = __builtin_shufflevector(__ret_35, __ret_35, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_35; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_p16(__p0_36, __p1_36, __p2_36, __p3_36) __extension__ ({ \
-  poly16x4_t __s0_36 = __p0_36; \
-  poly16x4_t __s2_36 = __p2_36; \
-  poly16x4_t __ret_36; \
-  __ret_36 = vset_lane_p16(vget_lane_p16(__s2_36, __p3_36), __s0_36, __p1_36); \
-  __ret_36; \
-})
-#else
-#define vcopy_lane_p16(__p0_37, __p1_37, __p2_37, __p3_37) __extension__ ({ \
-  poly16x4_t __s0_37 = __p0_37; \
-  poly16x4_t __s2_37 = __p2_37; \
-  poly16x4_t __rev0_37;  __rev0_37 = __builtin_shufflevector(__s0_37, __s0_37, 3, 2, 1, 0); \
-  poly16x4_t __rev2_37;  __rev2_37 = __builtin_shufflevector(__s2_37, __s2_37, 3, 2, 1, 0); \
-  poly16x4_t __ret_37; \
-  __ret_37 = __noswap_vset_lane_p16(__noswap_vget_lane_p16(__rev2_37, __p3_37), __rev0_37, __p1_37); \
-  __ret_37 = __builtin_shufflevector(__ret_37, __ret_37, 3, 2, 1, 0); \
-  __ret_37; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_u8(__p0_38, __p1_38, __p2_38, __p3_38) __extension__ ({ \
-  uint8x8_t __s0_38 = __p0_38; \
-  uint8x8_t __s2_38 = __p2_38; \
-  uint8x8_t __ret_38; \
-  __ret_38 = vset_lane_u8(vget_lane_u8(__s2_38, __p3_38), __s0_38, __p1_38); \
-  __ret_38; \
-})
-#else
-#define vcopy_lane_u8(__p0_39, __p1_39, __p2_39, __p3_39) __extension__ ({ \
-  uint8x8_t __s0_39 = __p0_39; \
-  uint8x8_t __s2_39 = __p2_39; \
-  uint8x8_t __rev0_39;  __rev0_39 = __builtin_shufflevector(__s0_39, __s0_39, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __rev2_39;  __rev2_39 = __builtin_shufflevector(__s2_39, __s2_39, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret_39; \
-  __ret_39 = __noswap_vset_lane_u8(__noswap_vget_lane_u8(__rev2_39, __p3_39), __rev0_39, __p1_39); \
-  __ret_39 = __builtin_shufflevector(__ret_39, __ret_39, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_39; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_u32(__p0_40, __p1_40, __p2_40, __p3_40) __extension__ ({ \
-  uint32x2_t __s0_40 = __p0_40; \
-  uint32x2_t __s2_40 = __p2_40; \
-  uint32x2_t __ret_40; \
-  __ret_40 = vset_lane_u32(vget_lane_u32(__s2_40, __p3_40), __s0_40, __p1_40); \
-  __ret_40; \
-})
-#else
-#define vcopy_lane_u32(__p0_41, __p1_41, __p2_41, __p3_41) __extension__ ({ \
-  uint32x2_t __s0_41 = __p0_41; \
-  uint32x2_t __s2_41 = __p2_41; \
-  uint32x2_t __rev0_41;  __rev0_41 = __builtin_shufflevector(__s0_41, __s0_41, 1, 0); \
-  uint32x2_t __rev2_41;  __rev2_41 = __builtin_shufflevector(__s2_41, __s2_41, 1, 0); \
-  uint32x2_t __ret_41; \
-  __ret_41 = __noswap_vset_lane_u32(__noswap_vget_lane_u32(__rev2_41, __p3_41), __rev0_41, __p1_41); \
-  __ret_41 = __builtin_shufflevector(__ret_41, __ret_41, 1, 0); \
-  __ret_41; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_u64(__p0_42, __p1_42, __p2_42, __p3_42) __extension__ ({ \
-  uint64x1_t __s0_42 = __p0_42; \
-  uint64x1_t __s2_42 = __p2_42; \
-  uint64x1_t __ret_42; \
-  __ret_42 = vset_lane_u64(vget_lane_u64(__s2_42, __p3_42), __s0_42, __p1_42); \
-  __ret_42; \
-})
-#else
-#define vcopy_lane_u64(__p0_43, __p1_43, __p2_43, __p3_43) __extension__ ({ \
-  uint64x1_t __s0_43 = __p0_43; \
-  uint64x1_t __s2_43 = __p2_43; \
-  uint64x1_t __ret_43; \
-  __ret_43 = __noswap_vset_lane_u64(__noswap_vget_lane_u64(__s2_43, __p3_43), __s0_43, __p1_43); \
-  __ret_43; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_u16(__p0_44, __p1_44, __p2_44, __p3_44) __extension__ ({ \
-  uint16x4_t __s0_44 = __p0_44; \
-  uint16x4_t __s2_44 = __p2_44; \
-  uint16x4_t __ret_44; \
-  __ret_44 = vset_lane_u16(vget_lane_u16(__s2_44, __p3_44), __s0_44, __p1_44); \
-  __ret_44; \
-})
-#else
-#define vcopy_lane_u16(__p0_45, __p1_45, __p2_45, __p3_45) __extension__ ({ \
-  uint16x4_t __s0_45 = __p0_45; \
-  uint16x4_t __s2_45 = __p2_45; \
-  uint16x4_t __rev0_45;  __rev0_45 = __builtin_shufflevector(__s0_45, __s0_45, 3, 2, 1, 0); \
-  uint16x4_t __rev2_45;  __rev2_45 = __builtin_shufflevector(__s2_45, __s2_45, 3, 2, 1, 0); \
-  uint16x4_t __ret_45; \
-  __ret_45 = __noswap_vset_lane_u16(__noswap_vget_lane_u16(__rev2_45, __p3_45), __rev0_45, __p1_45); \
-  __ret_45 = __builtin_shufflevector(__ret_45, __ret_45, 3, 2, 1, 0); \
-  __ret_45; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s8(__p0_46, __p1_46, __p2_46, __p3_46) __extension__ ({ \
-  int8x8_t __s0_46 = __p0_46; \
-  int8x8_t __s2_46 = __p2_46; \
-  int8x8_t __ret_46; \
-  __ret_46 = vset_lane_s8(vget_lane_s8(__s2_46, __p3_46), __s0_46, __p1_46); \
-  __ret_46; \
-})
-#else
-#define vcopy_lane_s8(__p0_47, __p1_47, __p2_47, __p3_47) __extension__ ({ \
-  int8x8_t __s0_47 = __p0_47; \
-  int8x8_t __s2_47 = __p2_47; \
-  int8x8_t __rev0_47;  __rev0_47 = __builtin_shufflevector(__s0_47, __s0_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __rev2_47;  __rev2_47 = __builtin_shufflevector(__s2_47, __s2_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_47; \
-  __ret_47 = __noswap_vset_lane_s8(__noswap_vget_lane_s8(__rev2_47, __p3_47), __rev0_47, __p1_47); \
-  __ret_47 = __builtin_shufflevector(__ret_47, __ret_47, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_47; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_f32(__p0_48, __p1_48, __p2_48, __p3_48) __extension__ ({ \
-  float32x2_t __s0_48 = __p0_48; \
-  float32x2_t __s2_48 = __p2_48; \
-  float32x2_t __ret_48; \
-  __ret_48 = vset_lane_f32(vget_lane_f32(__s2_48, __p3_48), __s0_48, __p1_48); \
-  __ret_48; \
-})
-#else
-#define vcopy_lane_f32(__p0_49, __p1_49, __p2_49, __p3_49) __extension__ ({ \
-  float32x2_t __s0_49 = __p0_49; \
-  float32x2_t __s2_49 = __p2_49; \
-  float32x2_t __rev0_49;  __rev0_49 = __builtin_shufflevector(__s0_49, __s0_49, 1, 0); \
-  float32x2_t __rev2_49;  __rev2_49 = __builtin_shufflevector(__s2_49, __s2_49, 1, 0); \
-  float32x2_t __ret_49; \
-  __ret_49 = __noswap_vset_lane_f32(__noswap_vget_lane_f32(__rev2_49, __p3_49), __rev0_49, __p1_49); \
-  __ret_49 = __builtin_shufflevector(__ret_49, __ret_49, 1, 0); \
-  __ret_49; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s32(__p0_50, __p1_50, __p2_50, __p3_50) __extension__ ({ \
-  int32x2_t __s0_50 = __p0_50; \
-  int32x2_t __s2_50 = __p2_50; \
-  int32x2_t __ret_50; \
-  __ret_50 = vset_lane_s32(vget_lane_s32(__s2_50, __p3_50), __s0_50, __p1_50); \
-  __ret_50; \
-})
-#else
-#define vcopy_lane_s32(__p0_51, __p1_51, __p2_51, __p3_51) __extension__ ({ \
-  int32x2_t __s0_51 = __p0_51; \
-  int32x2_t __s2_51 = __p2_51; \
-  int32x2_t __rev0_51;  __rev0_51 = __builtin_shufflevector(__s0_51, __s0_51, 1, 0); \
-  int32x2_t __rev2_51;  __rev2_51 = __builtin_shufflevector(__s2_51, __s2_51, 1, 0); \
-  int32x2_t __ret_51; \
-  __ret_51 = __noswap_vset_lane_s32(__noswap_vget_lane_s32(__rev2_51, __p3_51), __rev0_51, __p1_51); \
-  __ret_51 = __builtin_shufflevector(__ret_51, __ret_51, 1, 0); \
-  __ret_51; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s64(__p0_52, __p1_52, __p2_52, __p3_52) __extension__ ({ \
-  int64x1_t __s0_52 = __p0_52; \
-  int64x1_t __s2_52 = __p2_52; \
-  int64x1_t __ret_52; \
-  __ret_52 = vset_lane_s64(vget_lane_s64(__s2_52, __p3_52), __s0_52, __p1_52); \
-  __ret_52; \
-})
-#else
-#define vcopy_lane_s64(__p0_53, __p1_53, __p2_53, __p3_53) __extension__ ({ \
-  int64x1_t __s0_53 = __p0_53; \
-  int64x1_t __s2_53 = __p2_53; \
-  int64x1_t __ret_53; \
-  __ret_53 = __noswap_vset_lane_s64(__noswap_vget_lane_s64(__s2_53, __p3_53), __s0_53, __p1_53); \
-  __ret_53; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_s16(__p0_54, __p1_54, __p2_54, __p3_54) __extension__ ({ \
-  int16x4_t __s0_54 = __p0_54; \
-  int16x4_t __s2_54 = __p2_54; \
-  int16x4_t __ret_54; \
-  __ret_54 = vset_lane_s16(vget_lane_s16(__s2_54, __p3_54), __s0_54, __p1_54); \
-  __ret_54; \
-})
-#else
-#define vcopy_lane_s16(__p0_55, __p1_55, __p2_55, __p3_55) __extension__ ({ \
-  int16x4_t __s0_55 = __p0_55; \
-  int16x4_t __s2_55 = __p2_55; \
-  int16x4_t __rev0_55;  __rev0_55 = __builtin_shufflevector(__s0_55, __s0_55, 3, 2, 1, 0); \
-  int16x4_t __rev2_55;  __rev2_55 = __builtin_shufflevector(__s2_55, __s2_55, 3, 2, 1, 0); \
-  int16x4_t __ret_55; \
-  __ret_55 = __noswap_vset_lane_s16(__noswap_vget_lane_s16(__rev2_55, __p3_55), __rev0_55, __p1_55); \
-  __ret_55 = __builtin_shufflevector(__ret_55, __ret_55, 3, 2, 1, 0); \
-  __ret_55; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p8(__p0_56, __p1_56, __p2_56, __p3_56) __extension__ ({ \
-  poly8x16_t __s0_56 = __p0_56; \
-  poly8x16_t __s2_56 = __p2_56; \
-  poly8x16_t __ret_56; \
-  __ret_56 = vsetq_lane_p8(vgetq_lane_p8(__s2_56, __p3_56), __s0_56, __p1_56); \
-  __ret_56; \
-})
-#else
-#define vcopyq_laneq_p8(__p0_57, __p1_57, __p2_57, __p3_57) __extension__ ({ \
-  poly8x16_t __s0_57 = __p0_57; \
-  poly8x16_t __s2_57 = __p2_57; \
-  poly8x16_t __rev0_57;  __rev0_57 = __builtin_shufflevector(__s0_57, __s0_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_57;  __rev2_57 = __builtin_shufflevector(__s2_57, __s2_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret_57; \
-  __ret_57 = __noswap_vsetq_lane_p8(__noswap_vgetq_lane_p8(__rev2_57, __p3_57), __rev0_57, __p1_57); \
-  __ret_57 = __builtin_shufflevector(__ret_57, __ret_57, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_57; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p16(__p0_58, __p1_58, __p2_58, __p3_58) __extension__ ({ \
-  poly16x8_t __s0_58 = __p0_58; \
-  poly16x8_t __s2_58 = __p2_58; \
-  poly16x8_t __ret_58; \
-  __ret_58 = vsetq_lane_p16(vgetq_lane_p16(__s2_58, __p3_58), __s0_58, __p1_58); \
-  __ret_58; \
-})
-#else
-#define vcopyq_laneq_p16(__p0_59, __p1_59, __p2_59, __p3_59) __extension__ ({ \
-  poly16x8_t __s0_59 = __p0_59; \
-  poly16x8_t __s2_59 = __p2_59; \
-  poly16x8_t __rev0_59;  __rev0_59 = __builtin_shufflevector(__s0_59, __s0_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __rev2_59;  __rev2_59 = __builtin_shufflevector(__s2_59, __s2_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret_59; \
-  __ret_59 = __noswap_vsetq_lane_p16(__noswap_vgetq_lane_p16(__rev2_59, __p3_59), __rev0_59, __p1_59); \
-  __ret_59 = __builtin_shufflevector(__ret_59, __ret_59, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_59; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u8(__p0_60, __p1_60, __p2_60, __p3_60) __extension__ ({ \
-  uint8x16_t __s0_60 = __p0_60; \
-  uint8x16_t __s2_60 = __p2_60; \
-  uint8x16_t __ret_60; \
-  __ret_60 = vsetq_lane_u8(vgetq_lane_u8(__s2_60, __p3_60), __s0_60, __p1_60); \
-  __ret_60; \
-})
-#else
-#define vcopyq_laneq_u8(__p0_61, __p1_61, __p2_61, __p3_61) __extension__ ({ \
-  uint8x16_t __s0_61 = __p0_61; \
-  uint8x16_t __s2_61 = __p2_61; \
-  uint8x16_t __rev0_61;  __rev0_61 = __builtin_shufflevector(__s0_61, __s0_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_61;  __rev2_61 = __builtin_shufflevector(__s2_61, __s2_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_61; \
-  __ret_61 = __noswap_vsetq_lane_u8(__noswap_vgetq_lane_u8(__rev2_61, __p3_61), __rev0_61, __p1_61); \
-  __ret_61 = __builtin_shufflevector(__ret_61, __ret_61, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_61; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u32(__p0_62, __p1_62, __p2_62, __p3_62) __extension__ ({ \
-  uint32x4_t __s0_62 = __p0_62; \
-  uint32x4_t __s2_62 = __p2_62; \
-  uint32x4_t __ret_62; \
-  __ret_62 = vsetq_lane_u32(vgetq_lane_u32(__s2_62, __p3_62), __s0_62, __p1_62); \
-  __ret_62; \
-})
-#else
-#define vcopyq_laneq_u32(__p0_63, __p1_63, __p2_63, __p3_63) __extension__ ({ \
-  uint32x4_t __s0_63 = __p0_63; \
-  uint32x4_t __s2_63 = __p2_63; \
-  uint32x4_t __rev0_63;  __rev0_63 = __builtin_shufflevector(__s0_63, __s0_63, 3, 2, 1, 0); \
-  uint32x4_t __rev2_63;  __rev2_63 = __builtin_shufflevector(__s2_63, __s2_63, 3, 2, 1, 0); \
-  uint32x4_t __ret_63; \
-  __ret_63 = __noswap_vsetq_lane_u32(__noswap_vgetq_lane_u32(__rev2_63, __p3_63), __rev0_63, __p1_63); \
-  __ret_63 = __builtin_shufflevector(__ret_63, __ret_63, 3, 2, 1, 0); \
-  __ret_63; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u64(__p0_64, __p1_64, __p2_64, __p3_64) __extension__ ({ \
-  uint64x2_t __s0_64 = __p0_64; \
-  uint64x2_t __s2_64 = __p2_64; \
-  uint64x2_t __ret_64; \
-  __ret_64 = vsetq_lane_u64(vgetq_lane_u64(__s2_64, __p3_64), __s0_64, __p1_64); \
-  __ret_64; \
-})
-#else
-#define vcopyq_laneq_u64(__p0_65, __p1_65, __p2_65, __p3_65) __extension__ ({ \
-  uint64x2_t __s0_65 = __p0_65; \
-  uint64x2_t __s2_65 = __p2_65; \
-  uint64x2_t __rev0_65;  __rev0_65 = __builtin_shufflevector(__s0_65, __s0_65, 1, 0); \
-  uint64x2_t __rev2_65;  __rev2_65 = __builtin_shufflevector(__s2_65, __s2_65, 1, 0); \
-  uint64x2_t __ret_65; \
-  __ret_65 = __noswap_vsetq_lane_u64(__noswap_vgetq_lane_u64(__rev2_65, __p3_65), __rev0_65, __p1_65); \
-  __ret_65 = __builtin_shufflevector(__ret_65, __ret_65, 1, 0); \
-  __ret_65; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_u16(__p0_66, __p1_66, __p2_66, __p3_66) __extension__ ({ \
-  uint16x8_t __s0_66 = __p0_66; \
-  uint16x8_t __s2_66 = __p2_66; \
-  uint16x8_t __ret_66; \
-  __ret_66 = vsetq_lane_u16(vgetq_lane_u16(__s2_66, __p3_66), __s0_66, __p1_66); \
-  __ret_66; \
-})
-#else
-#define vcopyq_laneq_u16(__p0_67, __p1_67, __p2_67, __p3_67) __extension__ ({ \
-  uint16x8_t __s0_67 = __p0_67; \
-  uint16x8_t __s2_67 = __p2_67; \
-  uint16x8_t __rev0_67;  __rev0_67 = __builtin_shufflevector(__s0_67, __s0_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2_67;  __rev2_67 = __builtin_shufflevector(__s2_67, __s2_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_67; \
-  __ret_67 = __noswap_vsetq_lane_u16(__noswap_vgetq_lane_u16(__rev2_67, __p3_67), __rev0_67, __p1_67); \
-  __ret_67 = __builtin_shufflevector(__ret_67, __ret_67, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_67; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s8(__p0_68, __p1_68, __p2_68, __p3_68) __extension__ ({ \
-  int8x16_t __s0_68 = __p0_68; \
-  int8x16_t __s2_68 = __p2_68; \
-  int8x16_t __ret_68; \
-  __ret_68 = vsetq_lane_s8(vgetq_lane_s8(__s2_68, __p3_68), __s0_68, __p1_68); \
-  __ret_68; \
-})
-#else
-#define vcopyq_laneq_s8(__p0_69, __p1_69, __p2_69, __p3_69) __extension__ ({ \
-  int8x16_t __s0_69 = __p0_69; \
-  int8x16_t __s2_69 = __p2_69; \
-  int8x16_t __rev0_69;  __rev0_69 = __builtin_shufflevector(__s0_69, __s0_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_69;  __rev2_69 = __builtin_shufflevector(__s2_69, __s2_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_69; \
-  __ret_69 = __noswap_vsetq_lane_s8(__noswap_vgetq_lane_s8(__rev2_69, __p3_69), __rev0_69, __p1_69); \
-  __ret_69 = __builtin_shufflevector(__ret_69, __ret_69, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_69; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f32(__p0_70, __p1_70, __p2_70, __p3_70) __extension__ ({ \
-  float32x4_t __s0_70 = __p0_70; \
-  float32x4_t __s2_70 = __p2_70; \
-  float32x4_t __ret_70; \
-  __ret_70 = vsetq_lane_f32(vgetq_lane_f32(__s2_70, __p3_70), __s0_70, __p1_70); \
-  __ret_70; \
-})
-#else
-#define vcopyq_laneq_f32(__p0_71, __p1_71, __p2_71, __p3_71) __extension__ ({ \
-  float32x4_t __s0_71 = __p0_71; \
-  float32x4_t __s2_71 = __p2_71; \
-  float32x4_t __rev0_71;  __rev0_71 = __builtin_shufflevector(__s0_71, __s0_71, 3, 2, 1, 0); \
-  float32x4_t __rev2_71;  __rev2_71 = __builtin_shufflevector(__s2_71, __s2_71, 3, 2, 1, 0); \
-  float32x4_t __ret_71; \
-  __ret_71 = __noswap_vsetq_lane_f32(__noswap_vgetq_lane_f32(__rev2_71, __p3_71), __rev0_71, __p1_71); \
-  __ret_71 = __builtin_shufflevector(__ret_71, __ret_71, 3, 2, 1, 0); \
-  __ret_71; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s32(__p0_72, __p1_72, __p2_72, __p3_72) __extension__ ({ \
-  int32x4_t __s0_72 = __p0_72; \
-  int32x4_t __s2_72 = __p2_72; \
-  int32x4_t __ret_72; \
-  __ret_72 = vsetq_lane_s32(vgetq_lane_s32(__s2_72, __p3_72), __s0_72, __p1_72); \
-  __ret_72; \
-})
-#else
-#define vcopyq_laneq_s32(__p0_73, __p1_73, __p2_73, __p3_73) __extension__ ({ \
-  int32x4_t __s0_73 = __p0_73; \
-  int32x4_t __s2_73 = __p2_73; \
-  int32x4_t __rev0_73;  __rev0_73 = __builtin_shufflevector(__s0_73, __s0_73, 3, 2, 1, 0); \
-  int32x4_t __rev2_73;  __rev2_73 = __builtin_shufflevector(__s2_73, __s2_73, 3, 2, 1, 0); \
-  int32x4_t __ret_73; \
-  __ret_73 = __noswap_vsetq_lane_s32(__noswap_vgetq_lane_s32(__rev2_73, __p3_73), __rev0_73, __p1_73); \
-  __ret_73 = __builtin_shufflevector(__ret_73, __ret_73, 3, 2, 1, 0); \
-  __ret_73; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s64(__p0_74, __p1_74, __p2_74, __p3_74) __extension__ ({ \
-  int64x2_t __s0_74 = __p0_74; \
-  int64x2_t __s2_74 = __p2_74; \
-  int64x2_t __ret_74; \
-  __ret_74 = vsetq_lane_s64(vgetq_lane_s64(__s2_74, __p3_74), __s0_74, __p1_74); \
-  __ret_74; \
-})
-#else
-#define vcopyq_laneq_s64(__p0_75, __p1_75, __p2_75, __p3_75) __extension__ ({ \
-  int64x2_t __s0_75 = __p0_75; \
-  int64x2_t __s2_75 = __p2_75; \
-  int64x2_t __rev0_75;  __rev0_75 = __builtin_shufflevector(__s0_75, __s0_75, 1, 0); \
-  int64x2_t __rev2_75;  __rev2_75 = __builtin_shufflevector(__s2_75, __s2_75, 1, 0); \
-  int64x2_t __ret_75; \
-  __ret_75 = __noswap_vsetq_lane_s64(__noswap_vgetq_lane_s64(__rev2_75, __p3_75), __rev0_75, __p1_75); \
-  __ret_75 = __builtin_shufflevector(__ret_75, __ret_75, 1, 0); \
-  __ret_75; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_s16(__p0_76, __p1_76, __p2_76, __p3_76) __extension__ ({ \
-  int16x8_t __s0_76 = __p0_76; \
-  int16x8_t __s2_76 = __p2_76; \
-  int16x8_t __ret_76; \
-  __ret_76 = vsetq_lane_s16(vgetq_lane_s16(__s2_76, __p3_76), __s0_76, __p1_76); \
-  __ret_76; \
-})
-#else
-#define vcopyq_laneq_s16(__p0_77, __p1_77, __p2_77, __p3_77) __extension__ ({ \
-  int16x8_t __s0_77 = __p0_77; \
-  int16x8_t __s2_77 = __p2_77; \
-  int16x8_t __rev0_77;  __rev0_77 = __builtin_shufflevector(__s0_77, __s0_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2_77;  __rev2_77 = __builtin_shufflevector(__s2_77, __s2_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_77; \
-  __ret_77 = __noswap_vsetq_lane_s16(__noswap_vgetq_lane_s16(__rev2_77, __p3_77), __rev0_77, __p1_77); \
-  __ret_77 = __builtin_shufflevector(__ret_77, __ret_77, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_77; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p8(__p0_78, __p1_78, __p2_78, __p3_78) __extension__ ({ \
-  poly8x8_t __s0_78 = __p0_78; \
-  poly8x16_t __s2_78 = __p2_78; \
-  poly8x8_t __ret_78; \
-  __ret_78 = vset_lane_p8(vgetq_lane_p8(__s2_78, __p3_78), __s0_78, __p1_78); \
-  __ret_78; \
-})
-#else
-#define vcopy_laneq_p8(__p0_79, __p1_79, __p2_79, __p3_79) __extension__ ({ \
-  poly8x8_t __s0_79 = __p0_79; \
-  poly8x16_t __s2_79 = __p2_79; \
-  poly8x8_t __rev0_79;  __rev0_79 = __builtin_shufflevector(__s0_79, __s0_79, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __rev2_79;  __rev2_79 = __builtin_shufflevector(__s2_79, __s2_79, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret_79; \
-  __ret_79 = __noswap_vset_lane_p8(__noswap_vgetq_lane_p8(__rev2_79, __p3_79), __rev0_79, __p1_79); \
-  __ret_79 = __builtin_shufflevector(__ret_79, __ret_79, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_79; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p16(__p0_80, __p1_80, __p2_80, __p3_80) __extension__ ({ \
-  poly16x4_t __s0_80 = __p0_80; \
-  poly16x8_t __s2_80 = __p2_80; \
-  poly16x4_t __ret_80; \
-  __ret_80 = vset_lane_p16(vgetq_lane_p16(__s2_80, __p3_80), __s0_80, __p1_80); \
-  __ret_80; \
-})
-#else
-#define vcopy_laneq_p16(__p0_81, __p1_81, __p2_81, __p3_81) __extension__ ({ \
-  poly16x4_t __s0_81 = __p0_81; \
-  poly16x8_t __s2_81 = __p2_81; \
-  poly16x4_t __rev0_81;  __rev0_81 = __builtin_shufflevector(__s0_81, __s0_81, 3, 2, 1, 0); \
-  poly16x8_t __rev2_81;  __rev2_81 = __builtin_shufflevector(__s2_81, __s2_81, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x4_t __ret_81; \
-  __ret_81 = __noswap_vset_lane_p16(__noswap_vgetq_lane_p16(__rev2_81, __p3_81), __rev0_81, __p1_81); \
-  __ret_81 = __builtin_shufflevector(__ret_81, __ret_81, 3, 2, 1, 0); \
-  __ret_81; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u8(__p0_82, __p1_82, __p2_82, __p3_82) __extension__ ({ \
-  uint8x8_t __s0_82 = __p0_82; \
-  uint8x16_t __s2_82 = __p2_82; \
-  uint8x8_t __ret_82; \
-  __ret_82 = vset_lane_u8(vgetq_lane_u8(__s2_82, __p3_82), __s0_82, __p1_82); \
-  __ret_82; \
-})
-#else
-#define vcopy_laneq_u8(__p0_83, __p1_83, __p2_83, __p3_83) __extension__ ({ \
-  uint8x8_t __s0_83 = __p0_83; \
-  uint8x16_t __s2_83 = __p2_83; \
-  uint8x8_t __rev0_83;  __rev0_83 = __builtin_shufflevector(__s0_83, __s0_83, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __rev2_83;  __rev2_83 = __builtin_shufflevector(__s2_83, __s2_83, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret_83; \
-  __ret_83 = __noswap_vset_lane_u8(__noswap_vgetq_lane_u8(__rev2_83, __p3_83), __rev0_83, __p1_83); \
-  __ret_83 = __builtin_shufflevector(__ret_83, __ret_83, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_83; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u32(__p0_84, __p1_84, __p2_84, __p3_84) __extension__ ({ \
-  uint32x2_t __s0_84 = __p0_84; \
-  uint32x4_t __s2_84 = __p2_84; \
-  uint32x2_t __ret_84; \
-  __ret_84 = vset_lane_u32(vgetq_lane_u32(__s2_84, __p3_84), __s0_84, __p1_84); \
-  __ret_84; \
-})
-#else
-#define vcopy_laneq_u32(__p0_85, __p1_85, __p2_85, __p3_85) __extension__ ({ \
-  uint32x2_t __s0_85 = __p0_85; \
-  uint32x4_t __s2_85 = __p2_85; \
-  uint32x2_t __rev0_85;  __rev0_85 = __builtin_shufflevector(__s0_85, __s0_85, 1, 0); \
-  uint32x4_t __rev2_85;  __rev2_85 = __builtin_shufflevector(__s2_85, __s2_85, 3, 2, 1, 0); \
-  uint32x2_t __ret_85; \
-  __ret_85 = __noswap_vset_lane_u32(__noswap_vgetq_lane_u32(__rev2_85, __p3_85), __rev0_85, __p1_85); \
-  __ret_85 = __builtin_shufflevector(__ret_85, __ret_85, 1, 0); \
-  __ret_85; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u64(__p0_86, __p1_86, __p2_86, __p3_86) __extension__ ({ \
-  uint64x1_t __s0_86 = __p0_86; \
-  uint64x2_t __s2_86 = __p2_86; \
-  uint64x1_t __ret_86; \
-  __ret_86 = vset_lane_u64(vgetq_lane_u64(__s2_86, __p3_86), __s0_86, __p1_86); \
-  __ret_86; \
-})
-#else
-#define vcopy_laneq_u64(__p0_87, __p1_87, __p2_87, __p3_87) __extension__ ({ \
-  uint64x1_t __s0_87 = __p0_87; \
-  uint64x2_t __s2_87 = __p2_87; \
-  uint64x2_t __rev2_87;  __rev2_87 = __builtin_shufflevector(__s2_87, __s2_87, 1, 0); \
-  uint64x1_t __ret_87; \
-  __ret_87 = __noswap_vset_lane_u64(__noswap_vgetq_lane_u64(__rev2_87, __p3_87), __s0_87, __p1_87); \
-  __ret_87; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_u16(__p0_88, __p1_88, __p2_88, __p3_88) __extension__ ({ \
-  uint16x4_t __s0_88 = __p0_88; \
-  uint16x8_t __s2_88 = __p2_88; \
-  uint16x4_t __ret_88; \
-  __ret_88 = vset_lane_u16(vgetq_lane_u16(__s2_88, __p3_88), __s0_88, __p1_88); \
-  __ret_88; \
-})
-#else
-#define vcopy_laneq_u16(__p0_89, __p1_89, __p2_89, __p3_89) __extension__ ({ \
-  uint16x4_t __s0_89 = __p0_89; \
-  uint16x8_t __s2_89 = __p2_89; \
-  uint16x4_t __rev0_89;  __rev0_89 = __builtin_shufflevector(__s0_89, __s0_89, 3, 2, 1, 0); \
-  uint16x8_t __rev2_89;  __rev2_89 = __builtin_shufflevector(__s2_89, __s2_89, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret_89; \
-  __ret_89 = __noswap_vset_lane_u16(__noswap_vgetq_lane_u16(__rev2_89, __p3_89), __rev0_89, __p1_89); \
-  __ret_89 = __builtin_shufflevector(__ret_89, __ret_89, 3, 2, 1, 0); \
-  __ret_89; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s8(__p0_90, __p1_90, __p2_90, __p3_90) __extension__ ({ \
-  int8x8_t __s0_90 = __p0_90; \
-  int8x16_t __s2_90 = __p2_90; \
-  int8x8_t __ret_90; \
-  __ret_90 = vset_lane_s8(vgetq_lane_s8(__s2_90, __p3_90), __s0_90, __p1_90); \
-  __ret_90; \
-})
-#else
-#define vcopy_laneq_s8(__p0_91, __p1_91, __p2_91, __p3_91) __extension__ ({ \
-  int8x8_t __s0_91 = __p0_91; \
-  int8x16_t __s2_91 = __p2_91; \
-  int8x8_t __rev0_91;  __rev0_91 = __builtin_shufflevector(__s0_91, __s0_91, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __rev2_91;  __rev2_91 = __builtin_shufflevector(__s2_91, __s2_91, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret_91; \
-  __ret_91 = __noswap_vset_lane_s8(__noswap_vgetq_lane_s8(__rev2_91, __p3_91), __rev0_91, __p1_91); \
-  __ret_91 = __builtin_shufflevector(__ret_91, __ret_91, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_91; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f32(__p0_92, __p1_92, __p2_92, __p3_92) __extension__ ({ \
-  float32x2_t __s0_92 = __p0_92; \
-  float32x4_t __s2_92 = __p2_92; \
-  float32x2_t __ret_92; \
-  __ret_92 = vset_lane_f32(vgetq_lane_f32(__s2_92, __p3_92), __s0_92, __p1_92); \
-  __ret_92; \
-})
-#else
-#define vcopy_laneq_f32(__p0_93, __p1_93, __p2_93, __p3_93) __extension__ ({ \
-  float32x2_t __s0_93 = __p0_93; \
-  float32x4_t __s2_93 = __p2_93; \
-  float32x2_t __rev0_93;  __rev0_93 = __builtin_shufflevector(__s0_93, __s0_93, 1, 0); \
-  float32x4_t __rev2_93;  __rev2_93 = __builtin_shufflevector(__s2_93, __s2_93, 3, 2, 1, 0); \
-  float32x2_t __ret_93; \
-  __ret_93 = __noswap_vset_lane_f32(__noswap_vgetq_lane_f32(__rev2_93, __p3_93), __rev0_93, __p1_93); \
-  __ret_93 = __builtin_shufflevector(__ret_93, __ret_93, 1, 0); \
-  __ret_93; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s32(__p0_94, __p1_94, __p2_94, __p3_94) __extension__ ({ \
-  int32x2_t __s0_94 = __p0_94; \
-  int32x4_t __s2_94 = __p2_94; \
-  int32x2_t __ret_94; \
-  __ret_94 = vset_lane_s32(vgetq_lane_s32(__s2_94, __p3_94), __s0_94, __p1_94); \
-  __ret_94; \
-})
-#else
-#define vcopy_laneq_s32(__p0_95, __p1_95, __p2_95, __p3_95) __extension__ ({ \
-  int32x2_t __s0_95 = __p0_95; \
-  int32x4_t __s2_95 = __p2_95; \
-  int32x2_t __rev0_95;  __rev0_95 = __builtin_shufflevector(__s0_95, __s0_95, 1, 0); \
-  int32x4_t __rev2_95;  __rev2_95 = __builtin_shufflevector(__s2_95, __s2_95, 3, 2, 1, 0); \
-  int32x2_t __ret_95; \
-  __ret_95 = __noswap_vset_lane_s32(__noswap_vgetq_lane_s32(__rev2_95, __p3_95), __rev0_95, __p1_95); \
-  __ret_95 = __builtin_shufflevector(__ret_95, __ret_95, 1, 0); \
-  __ret_95; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s64(__p0_96, __p1_96, __p2_96, __p3_96) __extension__ ({ \
-  int64x1_t __s0_96 = __p0_96; \
-  int64x2_t __s2_96 = __p2_96; \
-  int64x1_t __ret_96; \
-  __ret_96 = vset_lane_s64(vgetq_lane_s64(__s2_96, __p3_96), __s0_96, __p1_96); \
-  __ret_96; \
-})
-#else
-#define vcopy_laneq_s64(__p0_97, __p1_97, __p2_97, __p3_97) __extension__ ({ \
-  int64x1_t __s0_97 = __p0_97; \
-  int64x2_t __s2_97 = __p2_97; \
-  int64x2_t __rev2_97;  __rev2_97 = __builtin_shufflevector(__s2_97, __s2_97, 1, 0); \
-  int64x1_t __ret_97; \
-  __ret_97 = __noswap_vset_lane_s64(__noswap_vgetq_lane_s64(__rev2_97, __p3_97), __s0_97, __p1_97); \
-  __ret_97; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_s16(__p0_98, __p1_98, __p2_98, __p3_98) __extension__ ({ \
-  int16x4_t __s0_98 = __p0_98; \
-  int16x8_t __s2_98 = __p2_98; \
-  int16x4_t __ret_98; \
-  __ret_98 = vset_lane_s16(vgetq_lane_s16(__s2_98, __p3_98), __s0_98, __p1_98); \
-  __ret_98; \
-})
-#else
-#define vcopy_laneq_s16(__p0_99, __p1_99, __p2_99, __p3_99) __extension__ ({ \
-  int16x4_t __s0_99 = __p0_99; \
-  int16x8_t __s2_99 = __p2_99; \
-  int16x4_t __rev0_99;  __rev0_99 = __builtin_shufflevector(__s0_99, __s0_99, 3, 2, 1, 0); \
-  int16x8_t __rev2_99;  __rev2_99 = __builtin_shufflevector(__s2_99, __s2_99, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret_99; \
-  __ret_99 = __noswap_vset_lane_s16(__noswap_vgetq_lane_s16(__rev2_99, __p3_99), __rev0_99, __p1_99); \
-  __ret_99 = __builtin_shufflevector(__ret_99, __ret_99, 3, 2, 1, 0); \
-  __ret_99; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vcreate_p64(uint64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vcreate_p64(uint64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vcreate_f64(uint64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#else
-__ai float64x1_t vcreate_f64(uint64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t)(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vcvts_f32_s32(int32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vcvts_f32_s32(int32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vcvts_f32_u32(uint32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vcvts_f32_u32(uint32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvts_f32_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vcvt_f32_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vcvt_f32_f64(float64x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvt_f32_f64((int8x16_t)__p0, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vcvtd_f64_s64(int64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
-  return __ret;
-}
-#else
-__ai float64_t vcvtd_f64_s64(int64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
-  return __ret;
-}
-#else
-__ai float64_t vcvtd_f64_u64(uint64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vcvtd_f64_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai float64x2_t vcvtq_f64_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai float64x2_t vcvtq_f64_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvtq_f64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai float64x1_t vcvt_f64_u64(uint64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai float64x1_t vcvt_f64_s64(int64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vcvt_f64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vcvt_f64_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float64x2_t __noswap_vcvt_f64_f32(float32x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vcvt_f64_f32((int8x8_t)__p0, 42);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
-  float16x8_t __ret;
-  __ret = vcombine_f16(__p0, vcvt_f16_f32(__p1));
-  return __ret;
-}
-#else
-__ai float16x8_t vcvt_high_f16_f32(float16x4_t __p0, float32x4_t __p1) {
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float16x8_t __ret;
-  __ret = __noswap_vcombine_f16(__rev0, __noswap_vcvt_f16_f32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
-  float32x4_t __ret;
-  __ret = vcvt_f32_f16(vget_high_f16(__p0));
-  return __ret;
-}
-#else
-__ai float32x4_t vcvt_high_f32_f16(float16x8_t __p0) {
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vcvt_f32_f16(__noswap_vget_high_f16(__rev0));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
-  float32x4_t __ret;
-  __ret = vcombine_f32(__p0, vcvt_f32_f64(__p1));
-  return __ret;
-}
-#else
-__ai float32x4_t vcvt_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvt_f32_f64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
-  float64x2_t __ret;
-  __ret = vcvt_f64_f32(vget_high_f32(__p0));
-  return __ret;
-}
-#else
-__ai float64x2_t vcvt_high_f64_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float64x2_t __ret;
-  __ret = __noswap_vcvt_f64_f32(__noswap_vget_high_f32(__rev0));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvts_n_f32_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvts_n_f32_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vcvts_n_f32_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vcvtq_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vcvtq_n_f64_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vcvt_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vcvt_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vcvt_n_f64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvtd_n_f64_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvtd_n_f64_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vcvtd_n_f64_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvts_n_s32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vcvts_n_s32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__s0, __p1, 35); \
-  __ret; \
-})
-#else
-#define vcvtq_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = (int64x2_t) __builtin_neon_vcvtq_n_s64_v((int8x16_t)__rev0, __p1, 35); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#else
-#define vcvt_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = (int64x1_t) __builtin_neon_vcvt_n_s64_v((int8x8_t)__s0, __p1, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvtd_n_s64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vcvtd_n_s64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvts_n_u32_f32(__p0, __p1) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vcvts_n_u32_f32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__s0, __p1, 51); \
-  __ret; \
-})
-#else
-#define vcvtq_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = (uint64x2_t) __builtin_neon_vcvtq_n_u64_v((int8x16_t)__rev0, __p1, 51); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#else
-#define vcvt_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = (uint64x1_t) __builtin_neon_vcvt_n_u64_v((int8x8_t)__s0, __p1, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vcvtd_n_u64_f64(__p0, __p1) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vcvtd_n_u64_f64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vcvts_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vcvts_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvts_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcvtd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcvtd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vcvtq_s64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vcvtq_s64_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vcvt_s64_f64(float64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vcvt_s64_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcvts_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcvts_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvts_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcvtd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__p0, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vcvtq_u64_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vcvtq_u64_v((int8x16_t)__rev0, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vcvt_u64_f64(float64x1_t __p0) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vcvt_u64_v((int8x8_t)__p0, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vcvtas_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vcvtas_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtas_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcvtad_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcvtad_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtad_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcvtas_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtas_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcvtad_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtad_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vcvtms_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vcvtms_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtms_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcvtmd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtmd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcvtms_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtms_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcvtmd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtmd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vcvtns_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vcvtns_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtns_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcvtnd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtnd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcvtns_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtns_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcvtnd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtnd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vcvtps_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vcvtps_s32_f32(float32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vcvtps_s32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vcvtpd_s64_f64(float64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vcvtpd_s64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vcvtps_u32_f32(float32_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vcvtps_u32_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vcvtpd_u64_f64(float64_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vcvtpd_u64_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vcvtxd_f32_f64(float64_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vcvtxd_f32_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float32x2_t vcvtx_f32_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vcvtx_f32_f64(float64x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vcvtx_f32_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
-  float32x4_t __ret;
-  __ret = vcombine_f32(__p0, vcvtx_f32_f64(__p1));
-  return __ret;
-}
-#else
-__ai float32x4_t vcvtx_high_f32_f64(float32x2_t __p0, float64x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vcombine_f32(__rev0, __noswap_vcvtx_f32_f64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float64x2_t vdivq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 / __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float32x4_t vdivq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __rev0 / __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float64x1_t vdiv_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __p0 / __p1;
-  return __ret;
-}
-#else
-__ai float32x2_t vdiv_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __rev0 / __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_lane_p8(__p0, __p1) __extension__ ({ \
-  poly8x8_t __s0 = __p0; \
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_lane_p16(__p0, __p1) __extension__ ({ \
-  poly16x4_t __s0 = __p0; \
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_lane_u8(__p0, __p1) __extension__ ({ \
-  uint8x8_t __s0 = __p0; \
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_lane_u32(__p0, __p1) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_lane_u64(__p0, __p1) __extension__ ({ \
-  uint64x1_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_lane_u16(__p0, __p1) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_lane_s8(__p0, __p1) __extension__ ({ \
-  int8x8_t __s0 = __p0; \
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vdupb_lane_i8((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vdups_lane_f32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_lane_f32(__p0, __p1) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vdups_lane_f32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_lane_s32(__p0, __p1) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vdups_lane_i32((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_lane_s64(__p0, __p1) __extension__ ({ \
-  int64x1_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_lane_s16(__p0, __p1) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vduph_lane_i16((int8x8_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_lane_f16(__p0, __p1) __extension__ ({ \
-  float16x4_t __s0 = __p0; \
-  float16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8_t __ret; \
-  __ret = (poly8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16_t __ret; \
-  __ret = (poly16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupb_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vdupb_laneq_i8((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vdupd_laneq_f64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vdups_laneq_f32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vdups_laneq_f32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdups_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vdups_laneq_i32((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vdupd_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vdupd_laneq_i64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vduph_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vduph_laneq_i16((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64x1_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_p8(__p0, __p1) __extension__ ({ \
-  poly8x16_t __s0 = __p0; \
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_p16(__p0, __p1) __extension__ ({ \
-  poly16x8_t __s0 = __p0; \
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdupq_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdupq_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_u8(__p0, __p1) __extension__ ({ \
-  uint8x16_t __s0 = __p0; \
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_u32(__p0, __p1) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_u64(__p0, __p1) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint64x1_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_u16(__p0, __p1) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x8_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_s8(__p0, __p1) __extension__ ({ \
-  int8x16_t __s0 = __p0; \
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x8_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x1_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_f32(__p0, __p1) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_f16(__p0, __p1) __extension__ ({ \
-  float16x8_t __s0 = __p0; \
-  float16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_s32(__p0, __p1) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_s64(__p0, __p1) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int64x1_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vdup_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __ret; \
-  __ret = __builtin_shufflevector(__s0, __s0, __p1, __p1, __p1, __p1); \
-  __ret; \
-})
-#else
-#define vdup_laneq_s16(__p0, __p1) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __builtin_shufflevector(__rev0, __rev0, __p1, __p1, __p1, __p1); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai poly64x1_t vdup_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai poly64x2_t vdupq_n_p64(poly64_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vdupq_n_f64(float64_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai float64x2_t vdupq_n_f64(float64_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vdup_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai float64x1_t vdup_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#else
-#define vext_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
-  __ret; \
-})
-#else
-#define vextq_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 42); \
-  __ret; \
-})
-#else
-#define vextq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vextq_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#else
-#define vext_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vext_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float64x2_t __noswap_vfmaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vfmaq_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 42);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-__ai float64x1_t __noswap_vfma_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vfma_v((int8x8_t)__p0, (int8x8_t)__p1, (int8x8_t)__p2, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmad_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_lane_f64(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmas_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_lane_f32(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
-  __ret; \
-})
-#else
-#define vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__s2, __p3, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 42); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
-  __ret; \
-})
-#else
-#define vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, __p3, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_lane_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x8_t)__s2, __p3, 41); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#else
-#define vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#define __noswap_vfma_lane_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
-  __ret; \
-})
-#else
-#define vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x8_t)__rev2, __p3, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfma_lane_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_lane_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x8_t)__s2, __p3, 9); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmad_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vfmad_laneq_f64(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#define __noswap_vfmas_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32_t __s0 = __p0; \
-  float32_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32_t __ret; \
-  __ret = (float32_t) __builtin_neon_vfmas_laneq_f32(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
-  __ret; \
-})
-#else
-#define vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 42); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
-  __ret; \
-})
-#else
-#define vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, __p3, 41); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfmaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = (float32x4_t) __builtin_neon_vfmaq_laneq_v((int8x16_t)__s0, (int8x16_t)__s1, (int8x16_t)__s2, __p3, 41); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
-  __ret; \
-})
-#else
-#define vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__rev2, __p3, 10); \
-  __ret; \
-})
-#define __noswap_vfma_laneq_f64(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __s2 = __p2; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
-  __ret; \
-})
-#else
-#define vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__rev0, (int8x8_t)__rev1, (int8x16_t)__rev2, __p3, 9); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vfma_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = (float32x2_t) __builtin_neon_vfma_laneq_v((int8x8_t)__s0, (int8x8_t)__s1, (int8x16_t)__s2, __p3, 9); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __ret;
-  __ret = vfmaq_f64(__p0, __p1, (float64x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai float64x2_t vfmaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __noswap_vfmaq_f64(__rev0, __rev1, (float64x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = vfma_f64(__p0, __p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#else
-__ai float64x1_t vfma_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, __p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = vfmaq_f64(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float64x2_t vfmsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float64x2_t __ret;
-  __ret = __noswap_vfmaq_f64(__rev0, -__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = vfma_f64(__p0, -__p1, __p2);
-  return __ret;
-}
-#else
-__ai float64x1_t vfms_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsd_lane_f64(__p0_100, __p1_100, __p2_100, __p3_100) __extension__ ({ \
-  float64_t __s0_100 = __p0_100; \
-  float64_t __s1_100 = __p1_100; \
-  float64x1_t __s2_100 = __p2_100; \
-  float64_t __ret_100; \
-  __ret_100 = vfmad_lane_f64(__s0_100, -__s1_100, __s2_100, __p3_100); \
-  __ret_100; \
-})
-#else
-#define vfmsd_lane_f64(__p0_101, __p1_101, __p2_101, __p3_101) __extension__ ({ \
-  float64_t __s0_101 = __p0_101; \
-  float64_t __s1_101 = __p1_101; \
-  float64x1_t __s2_101 = __p2_101; \
-  float64_t __ret_101; \
-  __ret_101 = __noswap_vfmad_lane_f64(__s0_101, -__s1_101, __s2_101, __p3_101); \
-  __ret_101; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmss_lane_f32(__p0_102, __p1_102, __p2_102, __p3_102) __extension__ ({ \
-  float32_t __s0_102 = __p0_102; \
-  float32_t __s1_102 = __p1_102; \
-  float32x2_t __s2_102 = __p2_102; \
-  float32_t __ret_102; \
-  __ret_102 = vfmas_lane_f32(__s0_102, -__s1_102, __s2_102, __p3_102); \
-  __ret_102; \
-})
-#else
-#define vfmss_lane_f32(__p0_103, __p1_103, __p2_103, __p3_103) __extension__ ({ \
-  float32_t __s0_103 = __p0_103; \
-  float32_t __s1_103 = __p1_103; \
-  float32x2_t __s2_103 = __p2_103; \
-  float32x2_t __rev2_103;  __rev2_103 = __builtin_shufflevector(__s2_103, __s2_103, 1, 0); \
-  float32_t __ret_103; \
-  __ret_103 = __noswap_vfmas_lane_f32(__s0_103, -__s1_103, __rev2_103, __p3_103); \
-  __ret_103; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f64(__p0_104, __p1_104, __p2_104, __p3_104) __extension__ ({ \
-  float64x2_t __s0_104 = __p0_104; \
-  float64x2_t __s1_104 = __p1_104; \
-  float64x1_t __s2_104 = __p2_104; \
-  float64x2_t __ret_104; \
-  __ret_104 = vfmaq_lane_f64(__s0_104, -__s1_104, __s2_104, __p3_104); \
-  __ret_104; \
-})
-#else
-#define vfmsq_lane_f64(__p0_105, __p1_105, __p2_105, __p3_105) __extension__ ({ \
-  float64x2_t __s0_105 = __p0_105; \
-  float64x2_t __s1_105 = __p1_105; \
-  float64x1_t __s2_105 = __p2_105; \
-  float64x2_t __rev0_105;  __rev0_105 = __builtin_shufflevector(__s0_105, __s0_105, 1, 0); \
-  float64x2_t __rev1_105;  __rev1_105 = __builtin_shufflevector(__s1_105, __s1_105, 1, 0); \
-  float64x2_t __ret_105; \
-  __ret_105 = __noswap_vfmaq_lane_f64(__rev0_105, -__rev1_105, __s2_105, __p3_105); \
-  __ret_105 = __builtin_shufflevector(__ret_105, __ret_105, 1, 0); \
-  __ret_105; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_lane_f32(__p0_106, __p1_106, __p2_106, __p3_106) __extension__ ({ \
-  float32x4_t __s0_106 = __p0_106; \
-  float32x4_t __s1_106 = __p1_106; \
-  float32x2_t __s2_106 = __p2_106; \
-  float32x4_t __ret_106; \
-  __ret_106 = vfmaq_lane_f32(__s0_106, -__s1_106, __s2_106, __p3_106); \
-  __ret_106; \
-})
-#else
-#define vfmsq_lane_f32(__p0_107, __p1_107, __p2_107, __p3_107) __extension__ ({ \
-  float32x4_t __s0_107 = __p0_107; \
-  float32x4_t __s1_107 = __p1_107; \
-  float32x2_t __s2_107 = __p2_107; \
-  float32x4_t __rev0_107;  __rev0_107 = __builtin_shufflevector(__s0_107, __s0_107, 3, 2, 1, 0); \
-  float32x4_t __rev1_107;  __rev1_107 = __builtin_shufflevector(__s1_107, __s1_107, 3, 2, 1, 0); \
-  float32x2_t __rev2_107;  __rev2_107 = __builtin_shufflevector(__s2_107, __s2_107, 1, 0); \
-  float32x4_t __ret_107; \
-  __ret_107 = __noswap_vfmaq_lane_f32(__rev0_107, -__rev1_107, __rev2_107, __p3_107); \
-  __ret_107 = __builtin_shufflevector(__ret_107, __ret_107, 3, 2, 1, 0); \
-  __ret_107; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f64(__p0_108, __p1_108, __p2_108, __p3_108) __extension__ ({ \
-  float64x1_t __s0_108 = __p0_108; \
-  float64x1_t __s1_108 = __p1_108; \
-  float64x1_t __s2_108 = __p2_108; \
-  float64x1_t __ret_108; \
-  __ret_108 = vfma_lane_f64(__s0_108, -__s1_108, __s2_108, __p3_108); \
-  __ret_108; \
-})
-#else
-#define vfms_lane_f64(__p0_109, __p1_109, __p2_109, __p3_109) __extension__ ({ \
-  float64x1_t __s0_109 = __p0_109; \
-  float64x1_t __s1_109 = __p1_109; \
-  float64x1_t __s2_109 = __p2_109; \
-  float64x1_t __ret_109; \
-  __ret_109 = __noswap_vfma_lane_f64(__s0_109, -__s1_109, __s2_109, __p3_109); \
-  __ret_109; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_lane_f32(__p0_110, __p1_110, __p2_110, __p3_110) __extension__ ({ \
-  float32x2_t __s0_110 = __p0_110; \
-  float32x2_t __s1_110 = __p1_110; \
-  float32x2_t __s2_110 = __p2_110; \
-  float32x2_t __ret_110; \
-  __ret_110 = vfma_lane_f32(__s0_110, -__s1_110, __s2_110, __p3_110); \
-  __ret_110; \
-})
-#else
-#define vfms_lane_f32(__p0_111, __p1_111, __p2_111, __p3_111) __extension__ ({ \
-  float32x2_t __s0_111 = __p0_111; \
-  float32x2_t __s1_111 = __p1_111; \
-  float32x2_t __s2_111 = __p2_111; \
-  float32x2_t __rev0_111;  __rev0_111 = __builtin_shufflevector(__s0_111, __s0_111, 1, 0); \
-  float32x2_t __rev1_111;  __rev1_111 = __builtin_shufflevector(__s1_111, __s1_111, 1, 0); \
-  float32x2_t __rev2_111;  __rev2_111 = __builtin_shufflevector(__s2_111, __s2_111, 1, 0); \
-  float32x2_t __ret_111; \
-  __ret_111 = __noswap_vfma_lane_f32(__rev0_111, -__rev1_111, __rev2_111, __p3_111); \
-  __ret_111 = __builtin_shufflevector(__ret_111, __ret_111, 1, 0); \
-  __ret_111; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsd_laneq_f64(__p0_112, __p1_112, __p2_112, __p3_112) __extension__ ({ \
-  float64_t __s0_112 = __p0_112; \
-  float64_t __s1_112 = __p1_112; \
-  float64x2_t __s2_112 = __p2_112; \
-  float64_t __ret_112; \
-  __ret_112 = vfmad_laneq_f64(__s0_112, -__s1_112, __s2_112, __p3_112); \
-  __ret_112; \
-})
-#else
-#define vfmsd_laneq_f64(__p0_113, __p1_113, __p2_113, __p3_113) __extension__ ({ \
-  float64_t __s0_113 = __p0_113; \
-  float64_t __s1_113 = __p1_113; \
-  float64x2_t __s2_113 = __p2_113; \
-  float64x2_t __rev2_113;  __rev2_113 = __builtin_shufflevector(__s2_113, __s2_113, 1, 0); \
-  float64_t __ret_113; \
-  __ret_113 = __noswap_vfmad_laneq_f64(__s0_113, -__s1_113, __rev2_113, __p3_113); \
-  __ret_113; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmss_laneq_f32(__p0_114, __p1_114, __p2_114, __p3_114) __extension__ ({ \
-  float32_t __s0_114 = __p0_114; \
-  float32_t __s1_114 = __p1_114; \
-  float32x4_t __s2_114 = __p2_114; \
-  float32_t __ret_114; \
-  __ret_114 = vfmas_laneq_f32(__s0_114, -__s1_114, __s2_114, __p3_114); \
-  __ret_114; \
-})
-#else
-#define vfmss_laneq_f32(__p0_115, __p1_115, __p2_115, __p3_115) __extension__ ({ \
-  float32_t __s0_115 = __p0_115; \
-  float32_t __s1_115 = __p1_115; \
-  float32x4_t __s2_115 = __p2_115; \
-  float32x4_t __rev2_115;  __rev2_115 = __builtin_shufflevector(__s2_115, __s2_115, 3, 2, 1, 0); \
-  float32_t __ret_115; \
-  __ret_115 = __noswap_vfmas_laneq_f32(__s0_115, -__s1_115, __rev2_115, __p3_115); \
-  __ret_115; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f64(__p0_116, __p1_116, __p2_116, __p3_116) __extension__ ({ \
-  float64x2_t __s0_116 = __p0_116; \
-  float64x2_t __s1_116 = __p1_116; \
-  float64x2_t __s2_116 = __p2_116; \
-  float64x2_t __ret_116; \
-  __ret_116 = vfmaq_laneq_f64(__s0_116, -__s1_116, __s2_116, __p3_116); \
-  __ret_116; \
-})
-#else
-#define vfmsq_laneq_f64(__p0_117, __p1_117, __p2_117, __p3_117) __extension__ ({ \
-  float64x2_t __s0_117 = __p0_117; \
-  float64x2_t __s1_117 = __p1_117; \
-  float64x2_t __s2_117 = __p2_117; \
-  float64x2_t __rev0_117;  __rev0_117 = __builtin_shufflevector(__s0_117, __s0_117, 1, 0); \
-  float64x2_t __rev1_117;  __rev1_117 = __builtin_shufflevector(__s1_117, __s1_117, 1, 0); \
-  float64x2_t __rev2_117;  __rev2_117 = __builtin_shufflevector(__s2_117, __s2_117, 1, 0); \
-  float64x2_t __ret_117; \
-  __ret_117 = __noswap_vfmaq_laneq_f64(__rev0_117, -__rev1_117, __rev2_117, __p3_117); \
-  __ret_117 = __builtin_shufflevector(__ret_117, __ret_117, 1, 0); \
-  __ret_117; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmsq_laneq_f32(__p0_118, __p1_118, __p2_118, __p3_118) __extension__ ({ \
-  float32x4_t __s0_118 = __p0_118; \
-  float32x4_t __s1_118 = __p1_118; \
-  float32x4_t __s2_118 = __p2_118; \
-  float32x4_t __ret_118; \
-  __ret_118 = vfmaq_laneq_f32(__s0_118, -__s1_118, __s2_118, __p3_118); \
-  __ret_118; \
-})
-#else
-#define vfmsq_laneq_f32(__p0_119, __p1_119, __p2_119, __p3_119) __extension__ ({ \
-  float32x4_t __s0_119 = __p0_119; \
-  float32x4_t __s1_119 = __p1_119; \
-  float32x4_t __s2_119 = __p2_119; \
-  float32x4_t __rev0_119;  __rev0_119 = __builtin_shufflevector(__s0_119, __s0_119, 3, 2, 1, 0); \
-  float32x4_t __rev1_119;  __rev1_119 = __builtin_shufflevector(__s1_119, __s1_119, 3, 2, 1, 0); \
-  float32x4_t __rev2_119;  __rev2_119 = __builtin_shufflevector(__s2_119, __s2_119, 3, 2, 1, 0); \
-  float32x4_t __ret_119; \
-  __ret_119 = __noswap_vfmaq_laneq_f32(__rev0_119, -__rev1_119, __rev2_119, __p3_119); \
-  __ret_119 = __builtin_shufflevector(__ret_119, __ret_119, 3, 2, 1, 0); \
-  __ret_119; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f64(__p0_120, __p1_120, __p2_120, __p3_120) __extension__ ({ \
-  float64x1_t __s0_120 = __p0_120; \
-  float64x1_t __s1_120 = __p1_120; \
-  float64x2_t __s2_120 = __p2_120; \
-  float64x1_t __ret_120; \
-  __ret_120 = vfma_laneq_f64(__s0_120, -__s1_120, __s2_120, __p3_120); \
-  __ret_120; \
-})
-#else
-#define vfms_laneq_f64(__p0_121, __p1_121, __p2_121, __p3_121) __extension__ ({ \
-  float64x1_t __s0_121 = __p0_121; \
-  float64x1_t __s1_121 = __p1_121; \
-  float64x2_t __s2_121 = __p2_121; \
-  float64x2_t __rev2_121;  __rev2_121 = __builtin_shufflevector(__s2_121, __s2_121, 1, 0); \
-  float64x1_t __ret_121; \
-  __ret_121 = __noswap_vfma_laneq_f64(__s0_121, -__s1_121, __rev2_121, __p3_121); \
-  __ret_121; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfms_laneq_f32(__p0_122, __p1_122, __p2_122, __p3_122) __extension__ ({ \
-  float32x2_t __s0_122 = __p0_122; \
-  float32x2_t __s1_122 = __p1_122; \
-  float32x4_t __s2_122 = __p2_122; \
-  float32x2_t __ret_122; \
-  __ret_122 = vfma_laneq_f32(__s0_122, -__s1_122, __s2_122, __p3_122); \
-  __ret_122; \
-})
-#else
-#define vfms_laneq_f32(__p0_123, __p1_123, __p2_123, __p3_123) __extension__ ({ \
-  float32x2_t __s0_123 = __p0_123; \
-  float32x2_t __s1_123 = __p1_123; \
-  float32x4_t __s2_123 = __p2_123; \
-  float32x2_t __rev0_123;  __rev0_123 = __builtin_shufflevector(__s0_123, __s0_123, 1, 0); \
-  float32x2_t __rev1_123;  __rev1_123 = __builtin_shufflevector(__s1_123, __s1_123, 1, 0); \
-  float32x4_t __rev2_123;  __rev2_123 = __builtin_shufflevector(__s2_123, __s2_123, 3, 2, 1, 0); \
-  float32x2_t __ret_123; \
-  __ret_123 = __noswap_vfma_laneq_f32(__rev0_123, -__rev1_123, __rev2_123, __p3_123); \
-  __ret_123 = __builtin_shufflevector(__ret_123, __ret_123, 1, 0); \
-  __ret_123; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __ret;
-  __ret = vfmaq_f64(__p0, -__p1, (float64x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai float64x2_t vfmsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __noswap_vfmaq_f64(__rev0, -__rev1, (float64x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __ret;
-  __ret = vfmaq_f32(__p0, -__p1, (float32x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai float32x4_t vfmsq_n_f32(float32x4_t __p0, float32x4_t __p1, float32_t __p2) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __noswap_vfmaq_f32(__rev0, -__rev1, (float32x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#else
-__ai float64x1_t vfms_n_f64(float64x1_t __p0, float64x1_t __p1, float64_t __p2) {
-  float64x1_t __ret;
-  __ret = __noswap_vfma_f64(__p0, -__p1, (float64x1_t) {__p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __ret;
-  __ret = vfma_f32(__p0, -__p1, (float32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai float32x2_t vfms_n_f32(float32x2_t __p0, float32x2_t __p1, float32_t __p2) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __noswap_vfma_f32(__rev0, -__rev1, (float32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1);
-  return __ret;
-}
-#else
-__ai poly64x1_t vget_high_p64(poly64x2_t __p0) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
-  return __ret;
-}
-__ai poly64x1_t __noswap_vget_high_p64(poly64x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vget_high_f64(float64x2_t __p0) {
-  float64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 1);
-  return __ret;
-}
-#else
-__ai float64x1_t vget_high_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vget_lane_i64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64_t __ret; \
-  __ret = (poly64_t) __builtin_neon_vgetq_lane_i64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vgetq_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__rev0, __p1); \
-  __ret; \
-})
-#define __noswap_vgetq_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vgetq_lane_f64((int8x16_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#else
-#define vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#define __noswap_vget_lane_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64_t __ret; \
-  __ret = (float64_t) __builtin_neon_vget_lane_f64((int8x8_t)__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
-  poly64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0);
-  return __ret;
-}
-#else
-__ai poly64x1_t vget_low_p64(poly64x2_t __p0) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vget_low_f64(float64x2_t __p0) {
-  float64x1_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p0, 0);
-  return __ret;
-}
-#else
-__ai float64x1_t vget_low_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x1_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev0, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
-  __ret; \
-})
-#else
-#define vld1_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p64(__p0) __extension__ ({ \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
-  __ret; \
-})
-#else
-#define vld1q_p64(__p0) __extension__ ({ \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_v(__p0, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f64(__p0) __extension__ ({ \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
-  __ret; \
-})
-#else
-#define vld1q_f64(__p0) __extension__ ({ \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_v(__p0, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
-  __ret; \
-})
-#else
-#define vld1_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
-  __ret; \
-})
-#else
-#define vld1_dup_p64(__p0) __extension__ ({ \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_dup_v(__p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_p64(__p0) __extension__ ({ \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
-  __ret; \
-})
-#else
-#define vld1q_dup_p64(__p0) __extension__ ({ \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_dup_v(__p0, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_dup_f64(__p0) __extension__ ({ \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
-  __ret; \
-})
-#else
-#define vld1q_dup_f64(__p0) __extension__ ({ \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_dup_v(__p0, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_dup_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
-  __ret; \
-})
-#else
-#define vld1_dup_f64(__p0) __extension__ ({ \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_dup_v(__p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#else
-#define vld1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
-  __ret; \
-})
-#else
-#define vld1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
-  __ret; \
-})
-#else
-#define vld1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vld1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#else
-#define vld1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vld1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p64_x2(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld1_p64_x2(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p64_x2(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld1q_p64_x2(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f64_x2(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld1q_f64_x2(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld1q_x2_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f64_x2(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld1_f64_x2(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld1_x2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p64_x3(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld1_p64_x3(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p64_x3(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld1q_p64_x3(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f64_x3(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld1q_f64_x3(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld1q_x3_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f64_x3(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld1_f64_x3(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld1_x3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_p64_x4(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld1_p64_x4(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_p64_x4(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld1q_p64_x4(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1q_f64_x4(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld1q_f64_x4(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld1q_x4_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld1_f64_x4(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld1_f64_x4(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld1_x4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld2_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_p64(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld2q_p64(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_u64(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld2q_u64(__p0) __extension__ ({ \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_f64(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld2q_f64(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_s64(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld2q_s64(__p0) __extension__ ({ \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld2_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld2_dup_p64(__p0) __extension__ ({ \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld2q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_dup_f64(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld2q_dup_f64(__p0) __extension__ ({ \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_dup_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_dup_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld2_dup_f64(__p0) __extension__ ({ \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-  __ret; \
-})
-#else
-#define vld2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  poly64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
-  __ret; \
-})
-#else
-#define vld2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  poly8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
-  __ret; \
-})
-#else
-#define vld2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  poly64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  poly64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
-  __ret; \
-})
-#else
-#define vld2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  uint8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
-  __ret; \
-})
-#else
-#define vld2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  uint64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  uint64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
-  __ret; \
-})
-#else
-#define vld2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  int8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 42); \
-  __ret; \
-})
-#else
-#define vld2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  float64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  float64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 35); \
-  __ret; \
-})
-#else
-#define vld2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  int64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  int64x2x2_t __ret; \
-  __builtin_neon_vld2q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __p2, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-  __ret; \
-})
-#else
-#define vld2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  uint64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
-  __ret; \
-})
-#else
-#define vld2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  float64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
-  __ret; \
-})
-#else
-#define vld2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  int64x1x2_t __ret; \
-  __builtin_neon_vld2_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld3_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_p64(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld3q_p64(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_u64(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld3q_u64(__p0) __extension__ ({ \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_f64(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld3q_f64(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_s64(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld3q_s64(__p0) __extension__ ({ \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld3_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld3_dup_p64(__p0) __extension__ ({ \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld3q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_dup_f64(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld3q_dup_f64(__p0) __extension__ ({ \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_dup_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_dup_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld3_dup_f64(__p0) __extension__ ({ \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-  __ret; \
-})
-#else
-#define vld3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  poly64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
-  __ret; \
-})
-#else
-#define vld3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  poly8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
-  __ret; \
-})
-#else
-#define vld3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  poly64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  poly64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
-  __ret; \
-})
-#else
-#define vld3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  uint8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
-  __ret; \
-})
-#else
-#define vld3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  uint64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  uint64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
-  __ret; \
-})
-#else
-#define vld3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  int8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 42); \
-  __ret; \
-})
-#else
-#define vld3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  float64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  float64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 35); \
-  __ret; \
-})
-#else
-#define vld3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  int64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  int64x2x3_t __ret; \
-  __builtin_neon_vld3q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-  __ret; \
-})
-#else
-#define vld3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  uint64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-  __ret; \
-})
-#else
-#define vld3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  float64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-  __ret; \
-})
-#else
-#define vld3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  int64x1x3_t __ret; \
-  __builtin_neon_vld3_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld4_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_p64(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld4q_p64(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_u64(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 51); \
-  __ret; \
-})
-#else
-#define vld4q_u64(__p0) __extension__ ({ \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_f64(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld4q_f64(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_s64(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 35); \
-  __ret; \
-})
-#else
-#define vld4q_s64(__p0) __extension__ ({ \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_v(&__ret, __p0, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld4_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#else
-#define vld4_dup_p64(__p0) __extension__ ({ \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
-  __ret; \
-})
-#else
-#define vld4q_dup_p64(__p0) __extension__ ({ \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_dup_f64(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
-  __ret; \
-})
-#else
-#define vld4q_dup_f64(__p0) __extension__ ({ \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_dup_v(&__ret, __p0, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_dup_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#else
-#define vld4_dup_f64(__p0) __extension__ ({ \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_dup_v(&__ret, __p0, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-  __ret; \
-})
-#else
-#define vld4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  poly64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
-  __ret; \
-})
-#else
-#define vld4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  poly8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  poly8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
-  __ret; \
-})
-#else
-#define vld4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  poly64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  poly64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
-  __ret; \
-})
-#else
-#define vld4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  uint8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
-  __ret; \
-})
-#else
-#define vld4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  uint64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  uint64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
-  __ret; \
-})
-#else
-#define vld4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  int8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 42); \
-  __ret; \
-})
-#else
-#define vld4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  float64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  float64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 42); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 35); \
-  __ret; \
-})
-#else
-#define vld4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  int64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  int64x2x4_t __ret; \
-  __builtin_neon_vld4q_lane_v(&__ret, __p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 35); \
- \
-  __ret.val[0] = __builtin_shufflevector(__ret.val[0], __ret.val[0], 1, 0); \
-  __ret.val[1] = __builtin_shufflevector(__ret.val[1], __ret.val[1], 1, 0); \
-  __ret.val[2] = __builtin_shufflevector(__ret.val[2], __ret.val[2], 1, 0); \
-  __ret.val[3] = __builtin_shufflevector(__ret.val[3], __ret.val[3], 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-  __ret; \
-})
-#else
-#define vld4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  uint64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-  __ret; \
-})
-#else
-#define vld4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  float64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-  __ret; \
-})
-#else
-#define vld4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  int64x1x4_t __ret; \
-  __builtin_neon_vld4_lane_v(&__ret, __p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vldrq_p128(__p0) __extension__ ({ \
-  poly128_t __ret; \
-  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
-  __ret; \
-})
-#else
-#define vldrq_p128(__p0) __extension__ ({ \
-  poly128_t __ret; \
-  __ret = (poly128_t) __builtin_neon_vldrq_p128(__p0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vmax_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmax_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmaxnmvq_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vmaxnmvq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmaxnmvq_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxnmvq_f32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vmaxnmvq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxnmvq_f32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vmaxnmv_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxnmv_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vmaxnmv_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxnmv_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vmaxvq_u8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vmaxvq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vmaxvq_u8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vmaxvq_u32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vmaxvq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vmaxvq_u32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vmaxvq_u16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vmaxvq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vmaxvq_u16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vmaxvq_s8(int8x16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vmaxvq_s8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vmaxvq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vmaxvq_s8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vmaxvq_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmaxvq_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vmaxvq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmaxvq_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vmaxvq_f32(float32x4_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxvq_f32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vmaxvq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxvq_f32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vmaxvq_s32(int32x4_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vmaxvq_s32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vmaxvq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vmaxvq_s32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vmaxvq_s16(int16x8_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vmaxvq_s16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vmaxvq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vmaxvq_s16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vmaxv_u8(uint8x8_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vmaxv_u8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vmaxv_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vmaxv_u8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vmaxv_u32(uint32x2_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vmaxv_u32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vmaxv_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vmaxv_u32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vmaxv_u16(uint16x4_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vmaxv_u16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vmaxv_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vmaxv_u16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vmaxv_s8(int8x8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vmaxv_s8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vmaxv_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vmaxv_s8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vmaxv_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxv_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vmaxv_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmaxv_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vmaxv_s32(int32x2_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vmaxv_s32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vmaxv_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vmaxv_s32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vmaxv_s16(int16x4_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vmaxv_s16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vmaxv_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vmaxv_s16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vminq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vmin_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmin_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vminnmvq_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vminnmvq_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vminnmvq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vminnmvq_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vminnmvq_f32(float32x4_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminnmvq_f32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vminnmvq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminnmvq_f32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vminnmv_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminnmv_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vminnmv_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminnmv_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vminvq_u8(uint8x16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vminvq_u8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vminvq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vminvq_u8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vminvq_u32(uint32x4_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vminvq_u32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vminvq_u32(uint32x4_t __p0) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vminvq_u32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vminvq_u16(uint16x8_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vminvq_u16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vminvq_u16(uint16x8_t __p0) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vminvq_u16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vminvq_s8(int8x16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vminvq_s8((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vminvq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vminvq_s8((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vminvq_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vminvq_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vminvq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vminvq_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vminvq_f32(float32x4_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminvq_f32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vminvq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminvq_f32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vminvq_s32(int32x4_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vminvq_s32((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vminvq_s32(int32x4_t __p0) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vminvq_s32((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vminvq_s16(int16x8_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vminvq_s16((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vminvq_s16(int16x8_t __p0) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vminvq_s16((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vminv_u8(uint8x8_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vminv_u8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vminv_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vminv_u8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vminv_u32(uint32x2_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vminv_u32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vminv_u32(uint32x2_t __p0) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vminv_u32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vminv_u16(uint16x4_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vminv_u16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vminv_u16(uint16x4_t __p0) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vminv_u16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vminv_s8(int8x8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vminv_s8((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int8_t vminv_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vminv_s8((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vminv_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminv_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vminv_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vminv_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vminv_s32(int32x2_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vminv_s32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int32_t vminv_s32(int32x2_t __p0) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vminv_s32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vminv_s16(int16x4_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vminv_s16((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai int16_t vminv_s16(int16x4_t __p0) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vminv_s16((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float64x2_t vmlaq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 + __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float64x1_t vmla_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 + __p1 * __p2;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x8_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlaq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlaq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmla_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = __s0 + __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmla_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 + __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __ret;
-  __ret = __p0 + __p1 * (float64x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai float64x2_t vmlaq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 + __rev1 * (float64x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 + vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 + vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 + vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 + vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 + vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 + vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float64x2_t vmlsq_f64(float64x2_t __p0, float64x2_t __p1, float64x2_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 - __rev1 * __rev2;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#else
-__ai float64x1_t vmls_f64(float64x1_t __p0, float64x1_t __p1, float64x1_t __p2) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1 * __p2;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x8_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmlsq_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_laneq_f32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __s2 = __p2; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmls_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __ret; \
-  __ret = __s0 - __s1 * __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3); \
-  __ret; \
-})
-#else
-#define vmls_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 - __rev1 * __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __ret;
-  __ret = __p0 - __p1 * (float64x2_t) {__p2, __p2};
-  return __ret;
-}
-#else
-__ai float64x2_t vmlsq_n_f64(float64x2_t __p0, float64x2_t __p1, float64_t __p2) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 - __rev1 * (float64x2_t) {__p2, __p2};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 - vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 - vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 - vmull_u32(vget_high_u32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u32(__noswap_vget_high_u32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - vmull_u16(vget_high_u16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u16(__noswap_vget_high_u16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 - vmull_s32(vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s32(__noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - vmull_s16(vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s16(__noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 - vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_laneq_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_laneq_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 - vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai poly64x1_t vmov_n_p64(poly64_t __p0) {
-  poly64x1_t __ret;
-  __ret = (poly64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai poly64x2_t vmovq_n_p64(poly64_t __p0) {
-  poly64x2_t __ret;
-  __ret = (poly64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmovq_n_f64(float64_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) {__p0, __p0};
-  return __ret;
-}
-#else
-__ai float64x2_t vmovq_n_f64(float64_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) {__p0, __p0};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmov_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#else
-__ai float64x1_t vmov_n_f64(float64_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) {__p0};
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_124) {
-  uint16x8_t __ret_124;
-  uint8x8_t __a1_124 = vget_high_u8(__p0_124);
-  __ret_124 = (uint16x8_t)(vshll_n_u8(__a1_124, 0));
-  return __ret_124;
-}
-#else
-__ai uint16x8_t vmovl_high_u8(uint8x16_t __p0_125) {
-  uint8x16_t __rev0_125;  __rev0_125 = __builtin_shufflevector(__p0_125, __p0_125, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret_125;
-  uint8x8_t __a1_125 = __noswap_vget_high_u8(__rev0_125);
-  __ret_125 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_125, 0));
-  __ret_125 = __builtin_shufflevector(__ret_125, __ret_125, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret_125;
-}
-__ai uint16x8_t __noswap_vmovl_high_u8(uint8x16_t __p0_126) {
-  uint16x8_t __ret_126;
-  uint8x8_t __a1_126 = __noswap_vget_high_u8(__p0_126);
-  __ret_126 = (uint16x8_t)(__noswap_vshll_n_u8(__a1_126, 0));
-  return __ret_126;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_127) {
-  uint64x2_t __ret_127;
-  uint32x2_t __a1_127 = vget_high_u32(__p0_127);
-  __ret_127 = (uint64x2_t)(vshll_n_u32(__a1_127, 0));
-  return __ret_127;
-}
-#else
-__ai uint64x2_t vmovl_high_u32(uint32x4_t __p0_128) {
-  uint32x4_t __rev0_128;  __rev0_128 = __builtin_shufflevector(__p0_128, __p0_128, 3, 2, 1, 0);
-  uint64x2_t __ret_128;
-  uint32x2_t __a1_128 = __noswap_vget_high_u32(__rev0_128);
-  __ret_128 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_128, 0));
-  __ret_128 = __builtin_shufflevector(__ret_128, __ret_128, 1, 0);
-  return __ret_128;
-}
-__ai uint64x2_t __noswap_vmovl_high_u32(uint32x4_t __p0_129) {
-  uint64x2_t __ret_129;
-  uint32x2_t __a1_129 = __noswap_vget_high_u32(__p0_129);
-  __ret_129 = (uint64x2_t)(__noswap_vshll_n_u32(__a1_129, 0));
-  return __ret_129;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_130) {
-  uint32x4_t __ret_130;
-  uint16x4_t __a1_130 = vget_high_u16(__p0_130);
-  __ret_130 = (uint32x4_t)(vshll_n_u16(__a1_130, 0));
-  return __ret_130;
-}
-#else
-__ai uint32x4_t vmovl_high_u16(uint16x8_t __p0_131) {
-  uint16x8_t __rev0_131;  __rev0_131 = __builtin_shufflevector(__p0_131, __p0_131, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret_131;
-  uint16x4_t __a1_131 = __noswap_vget_high_u16(__rev0_131);
-  __ret_131 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_131, 0));
-  __ret_131 = __builtin_shufflevector(__ret_131, __ret_131, 3, 2, 1, 0);
-  return __ret_131;
-}
-__ai uint32x4_t __noswap_vmovl_high_u16(uint16x8_t __p0_132) {
-  uint32x4_t __ret_132;
-  uint16x4_t __a1_132 = __noswap_vget_high_u16(__p0_132);
-  __ret_132 = (uint32x4_t)(__noswap_vshll_n_u16(__a1_132, 0));
-  return __ret_132;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_133) {
-  int16x8_t __ret_133;
-  int8x8_t __a1_133 = vget_high_s8(__p0_133);
-  __ret_133 = (int16x8_t)(vshll_n_s8(__a1_133, 0));
-  return __ret_133;
-}
-#else
-__ai int16x8_t vmovl_high_s8(int8x16_t __p0_134) {
-  int8x16_t __rev0_134;  __rev0_134 = __builtin_shufflevector(__p0_134, __p0_134, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret_134;
-  int8x8_t __a1_134 = __noswap_vget_high_s8(__rev0_134);
-  __ret_134 = (int16x8_t)(__noswap_vshll_n_s8(__a1_134, 0));
-  __ret_134 = __builtin_shufflevector(__ret_134, __ret_134, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret_134;
-}
-__ai int16x8_t __noswap_vmovl_high_s8(int8x16_t __p0_135) {
-  int16x8_t __ret_135;
-  int8x8_t __a1_135 = __noswap_vget_high_s8(__p0_135);
-  __ret_135 = (int16x8_t)(__noswap_vshll_n_s8(__a1_135, 0));
-  return __ret_135;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_136) {
-  int64x2_t __ret_136;
-  int32x2_t __a1_136 = vget_high_s32(__p0_136);
-  __ret_136 = (int64x2_t)(vshll_n_s32(__a1_136, 0));
-  return __ret_136;
-}
-#else
-__ai int64x2_t vmovl_high_s32(int32x4_t __p0_137) {
-  int32x4_t __rev0_137;  __rev0_137 = __builtin_shufflevector(__p0_137, __p0_137, 3, 2, 1, 0);
-  int64x2_t __ret_137;
-  int32x2_t __a1_137 = __noswap_vget_high_s32(__rev0_137);
-  __ret_137 = (int64x2_t)(__noswap_vshll_n_s32(__a1_137, 0));
-  __ret_137 = __builtin_shufflevector(__ret_137, __ret_137, 1, 0);
-  return __ret_137;
-}
-__ai int64x2_t __noswap_vmovl_high_s32(int32x4_t __p0_138) {
-  int64x2_t __ret_138;
-  int32x2_t __a1_138 = __noswap_vget_high_s32(__p0_138);
-  __ret_138 = (int64x2_t)(__noswap_vshll_n_s32(__a1_138, 0));
-  return __ret_138;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_139) {
-  int32x4_t __ret_139;
-  int16x4_t __a1_139 = vget_high_s16(__p0_139);
-  __ret_139 = (int32x4_t)(vshll_n_s16(__a1_139, 0));
-  return __ret_139;
-}
-#else
-__ai int32x4_t vmovl_high_s16(int16x8_t __p0_140) {
-  int16x8_t __rev0_140;  __rev0_140 = __builtin_shufflevector(__p0_140, __p0_140, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret_140;
-  int16x4_t __a1_140 = __noswap_vget_high_s16(__rev0_140);
-  __ret_140 = (int32x4_t)(__noswap_vshll_n_s16(__a1_140, 0));
-  __ret_140 = __builtin_shufflevector(__ret_140, __ret_140, 3, 2, 1, 0);
-  return __ret_140;
-}
-__ai int32x4_t __noswap_vmovl_high_s16(int16x8_t __p0_141) {
-  int32x4_t __ret_141;
-  int16x4_t __a1_141 = __noswap_vget_high_s16(__p0_141);
-  __ret_141 = (int32x4_t)(__noswap_vshll_n_s16(__a1_141, 0));
-  return __ret_141;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vmovn_u32(__p1));
-  return __ret;
-}
-#else
-__ai uint16x8_t vmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vmovn_u32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vmovn_u64(__p1));
-  return __ret;
-}
-#else
-__ai uint32x4_t vmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vmovn_u64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vmovn_u16(__p1));
-  return __ret;
-}
-#else
-__ai uint8x16_t vmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vmovn_u16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vmovn_s32(__p1));
-  return __ret;
-}
-#else
-__ai int16x8_t vmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vmovn_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vmovn_s64(__p1));
-  return __ret;
-}
-#else
-__ai int32x4_t vmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vmovn_s64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vmovn_s16(__p1));
-  return __ret;
-}
-#else
-__ai int8x16_t vmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vmovn_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float64x2_t vmulq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 * __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#else
-__ai float64x1_t vmul_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 * __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmuld_lane_f64(__p0_142, __p1_142, __p2_142) __extension__ ({ \
-  float64_t __s0_142 = __p0_142; \
-  float64x1_t __s1_142 = __p1_142; \
-  float64_t __ret_142; \
-  __ret_142 = __s0_142 * vget_lane_f64(__s1_142, __p2_142); \
-  __ret_142; \
-})
-#else
-#define vmuld_lane_f64(__p0_143, __p1_143, __p2_143) __extension__ ({ \
-  float64_t __s0_143 = __p0_143; \
-  float64x1_t __s1_143 = __p1_143; \
-  float64_t __ret_143; \
-  __ret_143 = __s0_143 * __noswap_vget_lane_f64(__s1_143, __p2_143); \
-  __ret_143; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmuls_lane_f32(__p0_144, __p1_144, __p2_144) __extension__ ({ \
-  float32_t __s0_144 = __p0_144; \
-  float32x2_t __s1_144 = __p1_144; \
-  float32_t __ret_144; \
-  __ret_144 = __s0_144 * vget_lane_f32(__s1_144, __p2_144); \
-  __ret_144; \
-})
-#else
-#define vmuls_lane_f32(__p0_145, __p1_145, __p2_145) __extension__ ({ \
-  float32_t __s0_145 = __p0_145; \
-  float32x2_t __s1_145 = __p1_145; \
-  float32x2_t __rev1_145;  __rev1_145 = __builtin_shufflevector(__s1_145, __s1_145, 1, 0); \
-  float32_t __ret_145; \
-  __ret_145 = __s0_145 * __noswap_vget_lane_f32(__rev1_145, __p2_145); \
-  __ret_145; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#else
-#define vmul_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_lane_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmuld_laneq_f64(__p0_146, __p1_146, __p2_146) __extension__ ({ \
-  float64_t __s0_146 = __p0_146; \
-  float64x2_t __s1_146 = __p1_146; \
-  float64_t __ret_146; \
-  __ret_146 = __s0_146 * vgetq_lane_f64(__s1_146, __p2_146); \
-  __ret_146; \
-})
-#else
-#define vmuld_laneq_f64(__p0_147, __p1_147, __p2_147) __extension__ ({ \
-  float64_t __s0_147 = __p0_147; \
-  float64x2_t __s1_147 = __p1_147; \
-  float64x2_t __rev1_147;  __rev1_147 = __builtin_shufflevector(__s1_147, __s1_147, 1, 0); \
-  float64_t __ret_147; \
-  __ret_147 = __s0_147 * __noswap_vgetq_lane_f64(__rev1_147, __p2_147); \
-  __ret_147; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmuls_laneq_f32(__p0_148, __p1_148, __p2_148) __extension__ ({ \
-  float32_t __s0_148 = __p0_148; \
-  float32x4_t __s1_148 = __p1_148; \
-  float32_t __ret_148; \
-  __ret_148 = __s0_148 * vgetq_lane_f32(__s1_148, __p2_148); \
-  __ret_148; \
-})
-#else
-#define vmuls_laneq_f32(__p0_149, __p1_149, __p2_149) __extension__ ({ \
-  float32_t __s0_149 = __p0_149; \
-  float32x4_t __s1_149 = __p1_149; \
-  float32x4_t __rev1_149;  __rev1_149 = __builtin_shufflevector(__s1_149, __s1_149, 3, 2, 1, 0); \
-  float32_t __ret_149; \
-  __ret_149 = __s0_149 * __noswap_vgetq_lane_f32(__rev1_149, __p2_149); \
-  __ret_149; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__s1, __p2, 10); \
-  __ret; \
-})
-#else
-#define vmul_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vmul_laneq_v((int8x8_t)__s0, (int8x16_t)__rev1, __p2, 10); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmulq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmul_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = __s0 * __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2); \
-  __ret; \
-})
-#else
-#define vmul_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __rev0 * __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
-  return __ret;
-}
-#else
-__ai float64x1_t vmul_n_f64(float64x1_t __p0, float64_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmul_n_f64((int8x8_t)__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
-  float64x2_t __ret;
-  __ret = __p0 * (float64x2_t) {__p1, __p1};
-  return __ret;
-}
-#else
-__ai float64x2_t vmulq_n_f64(float64x2_t __p0, float64_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 * (float64x2_t) {__p1, __p1};
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai poly128_t vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-__ai poly128_t __noswap_vmull_p64(poly64_t __p0, poly64_t __p1) {
-  poly128_t __ret;
-  __ret = (poly128_t) __builtin_neon_vmull_p64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly16x8_t __ret;
-  __ret = vmull_p8(vget_high_p8(__p0), vget_high_p8(__p1));
-  return __ret;
-}
-#else
-__ai poly16x8_t vmull_high_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __noswap_vmull_p8(__noswap_vget_high_p8(__rev0), __noswap_vget_high_p8(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = vmull_u8(vget_high_u8(__p0), vget_high_u8(__p1));
-  return __ret;
-}
-#else
-__ai uint16x8_t vmull_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmull_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmull_u32(vget_high_u32(__p0), vget_high_u32(__p1));
-  return __ret;
-}
-#else
-__ai uint64x2_t vmull_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmull_u16(vget_high_u16(__p0), vget_high_u16(__p1));
-  return __ret;
-}
-#else
-__ai uint32x4_t vmull_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = vmull_s8(vget_high_s8(__p0), vget_high_s8(__p1));
-  return __ret;
-}
-#else
-__ai int16x8_t vmull_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmull_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = vmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
-  return __ret;
-}
-#else
-__ai int64x2_t vmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = vmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
-  return __ret;
-}
-#else
-__ai int32x4_t vmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly128_t __ret;
-  __ret = vmull_p64((poly64_t)(vget_high_p64(__p0)), (poly64_t)(vget_high_p64(__p1)));
-  return __ret;
-}
-#else
-__ai poly128_t vmull_high_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly128_t __ret;
-  __ret = __noswap_vmull_p64((poly64_t)(__noswap_vget_high_p64(__rev0)), (poly64_t)(__noswap_vget_high_p64(__rev1)));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = vmull_u32(vget_high_u32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_lane_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = vmull_u16(vget_high_u16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_lane_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = vmull_u32(vget_high_u32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __noswap_vmull_u32(__noswap_vget_high_u32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = vmull_u16(vget_high_u16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x8_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __noswap_vmull_u16(__noswap_vget_high_u16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmull_n_u32(vget_high_u32(__p0), __p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmull_high_n_u32(uint32x4_t __p0, uint32_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmull_n_u32(__noswap_vget_high_u32(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmull_n_u16(vget_high_u16(__p0), __p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmull_high_n_u16(uint16x8_t __p0, uint16_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmull_n_u16(__noswap_vget_high_u16(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = vmull_n_s32(vget_high_s32(__p0), __p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = vmull_n_s16(vget_high_s16(__p0), __p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint64x2_t __ret; \
-  __ret = vmull_u32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_laneq_u32(__p0, __p1, __p2) __extension__ ({ \
-  uint32x2_t __s0 = __p0; \
-  uint32x4_t __s1 = __p1; \
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __noswap_vmull_u32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint32x4_t __ret; \
-  __ret = vmull_u16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_laneq_u16(__p0, __p1, __p2) __extension__ ({ \
-  uint16x4_t __s0 = __p0; \
-  uint16x8_t __s1 = __p1; \
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __noswap_vmull_u16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float64x2_t __noswap_vmulxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai float32x4_t __noswap_vmulxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vmulxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vmulx_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai float32x2_t __noswap_vmulx_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vmulx_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float64_t vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-__ai float64_t __noswap_vmulxd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vmulxd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float32_t vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-__ai float32_t __noswap_vmulxs_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vmulxs_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxd_lane_f64(__p0_150, __p1_150, __p2_150) __extension__ ({ \
-  float64_t __s0_150 = __p0_150; \
-  float64x1_t __s1_150 = __p1_150; \
-  float64_t __ret_150; \
-  __ret_150 = vmulxd_f64(__s0_150, vget_lane_f64(__s1_150, __p2_150)); \
-  __ret_150; \
-})
-#else
-#define vmulxd_lane_f64(__p0_151, __p1_151, __p2_151) __extension__ ({ \
-  float64_t __s0_151 = __p0_151; \
-  float64x1_t __s1_151 = __p1_151; \
-  float64_t __ret_151; \
-  __ret_151 = __noswap_vmulxd_f64(__s0_151, __noswap_vget_lane_f64(__s1_151, __p2_151)); \
-  __ret_151; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxs_lane_f32(__p0_152, __p1_152, __p2_152) __extension__ ({ \
-  float32_t __s0_152 = __p0_152; \
-  float32x2_t __s1_152 = __p1_152; \
-  float32_t __ret_152; \
-  __ret_152 = vmulxs_f32(__s0_152, vget_lane_f32(__s1_152, __p2_152)); \
-  __ret_152; \
-})
-#else
-#define vmulxs_lane_f32(__p0_153, __p1_153, __p2_153) __extension__ ({ \
-  float32_t __s0_153 = __p0_153; \
-  float32x2_t __s1_153 = __p1_153; \
-  float32x2_t __rev1_153;  __rev1_153 = __builtin_shufflevector(__s1_153, __s1_153, 1, 0); \
-  float32_t __ret_153; \
-  __ret_153 = __noswap_vmulxs_f32(__s0_153, __noswap_vget_lane_f32(__rev1_153, __p2_153)); \
-  __ret_153; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = vmulxq_f64(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __ret; \
-  __ret = __noswap_vmulxq_f64(__rev0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = vmulxq_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __noswap_vmulxq_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = vmulx_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulx_lane_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x2_t __s1 = __p1; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __noswap_vmulx_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxd_laneq_f64(__p0_154, __p1_154, __p2_154) __extension__ ({ \
-  float64_t __s0_154 = __p0_154; \
-  float64x2_t __s1_154 = __p1_154; \
-  float64_t __ret_154; \
-  __ret_154 = vmulxd_f64(__s0_154, vgetq_lane_f64(__s1_154, __p2_154)); \
-  __ret_154; \
-})
-#else
-#define vmulxd_laneq_f64(__p0_155, __p1_155, __p2_155) __extension__ ({ \
-  float64_t __s0_155 = __p0_155; \
-  float64x2_t __s1_155 = __p1_155; \
-  float64x2_t __rev1_155;  __rev1_155 = __builtin_shufflevector(__s1_155, __s1_155, 1, 0); \
-  float64_t __ret_155; \
-  __ret_155 = __noswap_vmulxd_f64(__s0_155, __noswap_vgetq_lane_f64(__rev1_155, __p2_155)); \
-  __ret_155; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxs_laneq_f32(__p0_156, __p1_156, __p2_156) __extension__ ({ \
-  float32_t __s0_156 = __p0_156; \
-  float32x4_t __s1_156 = __p1_156; \
-  float32_t __ret_156; \
-  __ret_156 = vmulxs_f32(__s0_156, vgetq_lane_f32(__s1_156, __p2_156)); \
-  __ret_156; \
-})
-#else
-#define vmulxs_laneq_f32(__p0_157, __p1_157, __p2_157) __extension__ ({ \
-  float32_t __s0_157 = __p0_157; \
-  float32x4_t __s1_157 = __p1_157; \
-  float32x4_t __rev1_157;  __rev1_157 = __builtin_shufflevector(__s1_157, __s1_157, 3, 2, 1, 0); \
-  float32_t __ret_157; \
-  __ret_157 = __noswap_vmulxs_f32(__s0_157, __noswap_vgetq_lane_f32(__rev1_157, __p2_157)); \
-  __ret_157; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = vmulxq_f64(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_laneq_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = __noswap_vmulxq_f64(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulxq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __ret; \
-  __ret = vmulxq_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulxq_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x4_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x4_t __ret; \
-  __ret = __noswap_vmulxq_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __ret; \
-  __ret = vmulx_f32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vmulx_laneq_f32(__p0, __p1, __p2) __extension__ ({ \
-  float32x2_t __s0 = __p0; \
-  float32x4_t __s1 = __p1; \
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  float32x2_t __ret; \
-  __ret = __noswap_vmulx_f32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vnegq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float64x2_t vnegq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vnegq_s64(int64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int64x2_t vnegq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = -__rev0;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vneg_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai float64x1_t vneg_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#else
-__ai int64x1_t vneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = -__p0;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vpaddq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vpaddq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vpaddq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vpaddq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vpaddq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vpaddq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vpaddq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vpaddq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vpaddq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vpaddq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vpaddd_u64(uint64x2_t __p0) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vpaddd_u64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai uint64_t vpaddd_u64(uint64x2_t __p0) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vpaddd_u64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vpaddd_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpaddd_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vpaddd_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpaddd_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vpaddd_s64(int64x2_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vpaddd_s64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai int64_t vpaddd_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vpaddd_s64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vpadds_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpadds_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vpadds_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpadds_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vpmaxq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vpmaxq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vpmaxq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vpmaxq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vpmaxq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vpmaxq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vpmaxq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vpmaxq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpmaxq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vpmaxqd_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpmaxqd_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vpmaxqd_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpmaxqd_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vpmaxs_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmaxs_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vpmaxs_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmaxs_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vpmaxnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vpmaxnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpmaxnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vpmaxnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpmaxnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vpmaxnmqd_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpmaxnmqd_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vpmaxnms_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmaxnms_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vpmaxnms_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmaxnms_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vpminq_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vpminq_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vpminq_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vpminq_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vpminq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vpminq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vpminq_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vpminq_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vpminq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vpminqd_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpminqd_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vpminqd_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpminqd_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vpmins_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmins_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vpmins_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpmins_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vpminnmq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__p0, (int8x16_t)__p1, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vpminnmq_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vpminnmq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__p0, (int8x8_t)__p1, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vpminnm_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vpminnm_v((int8x8_t)__rev0, (int8x8_t)__rev1, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vpminnmqd_f64(float64x2_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpminnmqd_f64((int8x16_t)__p0);
-  return __ret;
-}
-#else
-__ai float64_t vpminnmqd_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vpminnmqd_f64((int8x16_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vpminnms_f32(float32x2_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpminnms_f32((int8x8_t)__p0);
-  return __ret;
-}
-#else
-__ai float32_t vpminnms_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vpminnms_f32((int8x8_t)__rev0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqabsq_s64(int64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqabsq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqabsq_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqabs_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqabs_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqabsb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
-  return __ret;
-}
-#else
-__ai int8_t vqabsb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqabsb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqabss_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vqabss_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqabss_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vqabsd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqabsd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqabsh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
-  return __ret;
-}
-#else
-__ai int16_t vqabsh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqabsh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqaddb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqadds_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqaddd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqaddh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqaddb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqadds_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqaddd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqaddh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#else
-__ai int64_t vqdmlals_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlals_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#else
-__ai int32_t vqdmlalh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlalh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __ret;
-  __ret = vqdmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __ret;
-  __ret = vqdmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlal_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlal_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlal_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlal_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = vqdmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = vqdmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlals_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlals_lane_s32(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlalh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlalh_lane_s16(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlals_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlals_laneq_s32(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlalh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlalh_laneq_s16(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlal_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlal_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlal_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlal_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlal_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#else
-__ai int64_t vqdmlsls_s32(int64_t __p0, int32_t __p1, int32_t __p2) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmlsls_s32(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#else
-__ai int32_t vqdmlslh_s16(int32_t __p0, int16_t __p1, int16_t __p2) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmlslh_s16(__p0, __p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __ret;
-  __ret = vqdmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __ret;
-  __ret = vqdmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlsl_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_high_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlsl_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_high_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlsl_s32(__s0, vget_high_s32(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_high_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlsl_s16(__s0, vget_high_s16(__s1), __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_high_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = vqdmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = vqdmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlsls_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlsls_lane_s32(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, (int8x8_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlslh_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlslh_lane_s16(__s0, __s1, (int8x8_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlsls_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqdmlsls_laneq_s32(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, (int8x16_t)__s2, __p3); \
-  __ret; \
-})
-#else
-#define vqdmlslh_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqdmlslh_laneq_s16(__s0, __s1, (int8x16_t)__rev2, __p3); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = vqdmlsl_s32(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_laneq_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmlsl_s32(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = vqdmlsl_s16(__s0, __s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vqdmlsl_laneq_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmlsl_s16(__rev0, __rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_lane_s32(__p0_158, __p1_158, __p2_158) __extension__ ({ \
-  int32_t __s0_158 = __p0_158; \
-  int32x2_t __s1_158 = __p1_158; \
-  int32_t __ret_158; \
-  __ret_158 = vqdmulhs_s32(__s0_158, vget_lane_s32(__s1_158, __p2_158)); \
-  __ret_158; \
-})
-#else
-#define vqdmulhs_lane_s32(__p0_159, __p1_159, __p2_159) __extension__ ({ \
-  int32_t __s0_159 = __p0_159; \
-  int32x2_t __s1_159 = __p1_159; \
-  int32x2_t __rev1_159;  __rev1_159 = __builtin_shufflevector(__s1_159, __s1_159, 1, 0); \
-  int32_t __ret_159; \
-  __ret_159 = __noswap_vqdmulhs_s32(__s0_159, __noswap_vget_lane_s32(__rev1_159, __p2_159)); \
-  __ret_159; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_lane_s16(__p0_160, __p1_160, __p2_160) __extension__ ({ \
-  int16_t __s0_160 = __p0_160; \
-  int16x4_t __s1_160 = __p1_160; \
-  int16_t __ret_160; \
-  __ret_160 = vqdmulhh_s16(__s0_160, vget_lane_s16(__s1_160, __p2_160)); \
-  __ret_160; \
-})
-#else
-#define vqdmulhh_lane_s16(__p0_161, __p1_161, __p2_161) __extension__ ({ \
-  int16_t __s0_161 = __p0_161; \
-  int16x4_t __s1_161 = __p1_161; \
-  int16x4_t __rev1_161;  __rev1_161 = __builtin_shufflevector(__s1_161, __s1_161, 3, 2, 1, 0); \
-  int16_t __ret_161; \
-  __ret_161 = __noswap_vqdmulhh_s16(__s0_161, __noswap_vget_lane_s16(__rev1_161, __p2_161)); \
-  __ret_161; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhs_laneq_s32(__p0_162, __p1_162, __p2_162) __extension__ ({ \
-  int32_t __s0_162 = __p0_162; \
-  int32x4_t __s1_162 = __p1_162; \
-  int32_t __ret_162; \
-  __ret_162 = vqdmulhs_s32(__s0_162, vgetq_lane_s32(__s1_162, __p2_162)); \
-  __ret_162; \
-})
-#else
-#define vqdmulhs_laneq_s32(__p0_163, __p1_163, __p2_163) __extension__ ({ \
-  int32_t __s0_163 = __p0_163; \
-  int32x4_t __s1_163 = __p1_163; \
-  int32x4_t __rev1_163;  __rev1_163 = __builtin_shufflevector(__s1_163, __s1_163, 3, 2, 1, 0); \
-  int32_t __ret_163; \
-  __ret_163 = __noswap_vqdmulhs_s32(__s0_163, __noswap_vgetq_lane_s32(__rev1_163, __p2_163)); \
-  __ret_163; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhh_laneq_s16(__p0_164, __p1_164, __p2_164) __extension__ ({ \
-  int16_t __s0_164 = __p0_164; \
-  int16x8_t __s1_164 = __p1_164; \
-  int16_t __ret_164; \
-  __ret_164 = vqdmulhh_s16(__s0_164, vgetq_lane_s16(__s1_164, __p2_164)); \
-  __ret_164; \
-})
-#else
-#define vqdmulhh_laneq_s16(__p0_165, __p1_165, __p2_165) __extension__ ({ \
-  int16_t __s0_165 = __p0_165; \
-  int16x8_t __s1_165 = __p1_165; \
-  int16x8_t __rev1_165;  __rev1_165 = __builtin_shufflevector(__s1_165, __s1_165, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_165; \
-  __ret_165 = __noswap_vqdmulhh_s16(__s0_165, __noswap_vgetq_lane_s16(__rev1_165, __p2_165)); \
-  __ret_165; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = vqdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = vqdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = vqdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-__ai int64_t __noswap_vqdmulls_s32(int32_t __p0, int32_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqdmulls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqdmullh_s16(int16_t __p0, int16_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqdmullh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = vqdmull_s32(vget_high_s32(__p0), vget_high_s32(__p1));
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmull_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = vqdmull_s16(vget_high_s16(__p0), vget_high_s16(__p1));
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmull_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vqdmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_high_lane_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_high_lane_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vqdmull_s32(vget_high_s32(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_high_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmull_s32(__noswap_vget_high_s32(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmull_s16(vget_high_s16(__s0), __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_high_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmull_s16(__noswap_vget_high_s16(__rev0), __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
-  int64x2_t __ret;
-  __ret = vqdmull_n_s32(vget_high_s32(__p0), __p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vqdmull_high_n_s32(int32x4_t __p0, int32_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vqdmull_n_s32(__noswap_vget_high_s32(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
-  int32x4_t __ret;
-  __ret = vqdmull_n_s16(vget_high_s16(__p0), __p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vqdmull_high_n_s16(int16x8_t __p0, int16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vqdmull_n_s16(__noswap_vget_high_s16(__rev0), __p1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulls_lane_s32(__p0_166, __p1_166, __p2_166) __extension__ ({ \
-  int32_t __s0_166 = __p0_166; \
-  int32x2_t __s1_166 = __p1_166; \
-  int64_t __ret_166; \
-  __ret_166 = vqdmulls_s32(__s0_166, vget_lane_s32(__s1_166, __p2_166)); \
-  __ret_166; \
-})
-#else
-#define vqdmulls_lane_s32(__p0_167, __p1_167, __p2_167) __extension__ ({ \
-  int32_t __s0_167 = __p0_167; \
-  int32x2_t __s1_167 = __p1_167; \
-  int32x2_t __rev1_167;  __rev1_167 = __builtin_shufflevector(__s1_167, __s1_167, 1, 0); \
-  int64_t __ret_167; \
-  __ret_167 = __noswap_vqdmulls_s32(__s0_167, __noswap_vget_lane_s32(__rev1_167, __p2_167)); \
-  __ret_167; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmullh_lane_s16(__p0_168, __p1_168, __p2_168) __extension__ ({ \
-  int16_t __s0_168 = __p0_168; \
-  int16x4_t __s1_168 = __p1_168; \
-  int32_t __ret_168; \
-  __ret_168 = vqdmullh_s16(__s0_168, vget_lane_s16(__s1_168, __p2_168)); \
-  __ret_168; \
-})
-#else
-#define vqdmullh_lane_s16(__p0_169, __p1_169, __p2_169) __extension__ ({ \
-  int16_t __s0_169 = __p0_169; \
-  int16x4_t __s1_169 = __p1_169; \
-  int16x4_t __rev1_169;  __rev1_169 = __builtin_shufflevector(__s1_169, __s1_169, 3, 2, 1, 0); \
-  int32_t __ret_169; \
-  __ret_169 = __noswap_vqdmullh_s16(__s0_169, __noswap_vget_lane_s16(__rev1_169, __p2_169)); \
-  __ret_169; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmulls_laneq_s32(__p0_170, __p1_170, __p2_170) __extension__ ({ \
-  int32_t __s0_170 = __p0_170; \
-  int32x4_t __s1_170 = __p1_170; \
-  int64_t __ret_170; \
-  __ret_170 = vqdmulls_s32(__s0_170, vgetq_lane_s32(__s1_170, __p2_170)); \
-  __ret_170; \
-})
-#else
-#define vqdmulls_laneq_s32(__p0_171, __p1_171, __p2_171) __extension__ ({ \
-  int32_t __s0_171 = __p0_171; \
-  int32x4_t __s1_171 = __p1_171; \
-  int32x4_t __rev1_171;  __rev1_171 = __builtin_shufflevector(__s1_171, __s1_171, 3, 2, 1, 0); \
-  int64_t __ret_171; \
-  __ret_171 = __noswap_vqdmulls_s32(__s0_171, __noswap_vgetq_lane_s32(__rev1_171, __p2_171)); \
-  __ret_171; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmullh_laneq_s16(__p0_172, __p1_172, __p2_172) __extension__ ({ \
-  int16_t __s0_172 = __p0_172; \
-  int16x8_t __s1_172 = __p1_172; \
-  int32_t __ret_172; \
-  __ret_172 = vqdmullh_s16(__s0_172, vgetq_lane_s16(__s1_172, __p2_172)); \
-  __ret_172; \
-})
-#else
-#define vqdmullh_laneq_s16(__p0_173, __p1_173, __p2_173) __extension__ ({ \
-  int16_t __s0_173 = __p0_173; \
-  int16x8_t __s1_173 = __p1_173; \
-  int16x8_t __rev1_173;  __rev1_173 = __builtin_shufflevector(__s1_173, __s1_173, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32_t __ret_173; \
-  __ret_173 = __noswap_vqdmullh_s16(__s0_173, __noswap_vgetq_lane_s16(__rev1_173, __p2_173)); \
-  __ret_173; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int64x2_t __ret; \
-  __ret = vqdmull_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __noswap_vqdmull_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqdmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqdmull_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqdmull_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqdmull_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqmovns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
-  return __ret;
-}
-#else
-__ai int16_t vqmovns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqmovnd_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vqmovnd_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovnd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqmovnh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
-  return __ret;
-}
-#else
-__ai int8_t vqmovnh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovnh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqmovns_u32(uint32_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
-  return __ret;
-}
-#else
-__ai uint16_t vqmovns_u32(uint32_t __p0) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqmovns_u32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqmovnd_u64(uint64_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
-  return __ret;
-}
-#else
-__ai uint32_t vqmovnd_u64(uint64_t __p0) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqmovnd_u64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqmovnh_u16(uint16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
-  return __ret;
-}
-#else
-__ai uint8_t vqmovnh_u16(uint16_t __p0) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqmovnh_u16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vqmovn_u32(__p1));
-  return __ret;
-}
-#else
-__ai uint16x8_t vqmovn_high_u32(uint16x4_t __p0, uint32x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vqmovn_u32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vqmovn_u64(__p1));
-  return __ret;
-}
-#else
-__ai uint32x4_t vqmovn_high_u64(uint32x2_t __p0, uint64x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vqmovn_u64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vqmovn_u16(__p1));
-  return __ret;
-}
-#else
-__ai uint8x16_t vqmovn_high_u16(uint8x8_t __p0, uint16x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vqmovn_u16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vqmovn_s32(__p1));
-  return __ret;
-}
-#else
-__ai int16x8_t vqmovn_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vqmovn_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vqmovn_s64(__p1));
-  return __ret;
-}
-#else
-__ai int32x4_t vqmovn_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vqmovn_s64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vqmovn_s16(__p1));
-  return __ret;
-}
-#else
-__ai int8x16_t vqmovn_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vqmovn_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqmovuns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
-  return __ret;
-}
-#else
-__ai int16_t vqmovuns_s32(int32_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqmovuns_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqmovund_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vqmovund_s64(int64_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqmovund_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqmovunh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
-  return __ret;
-}
-#else
-__ai int8_t vqmovunh_s16(int16_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqmovunh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16((uint16x4_t)(__p0), vqmovun_s32(__p1));
-  return __ret;
-}
-#else
-__ai uint16x8_t vqmovun_high_s32(int16x4_t __p0, int32x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16((uint16x4_t)(__rev0), __noswap_vqmovun_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32((uint32x2_t)(__p0), vqmovun_s64(__p1));
-  return __ret;
-}
-#else
-__ai uint32x4_t vqmovun_high_s64(int32x2_t __p0, int64x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32((uint32x2_t)(__rev0), __noswap_vqmovun_s64(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8((uint8x8_t)(__p0), vqmovun_s16(__p1));
-  return __ret;
-}
-#else
-__ai uint8x16_t vqmovun_high_s16(int8x8_t __p0, int16x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8((uint8x8_t)(__rev0), __noswap_vqmovun_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vqnegq_s64(int64x2_t __p0) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__p0, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vqnegq_s64(int64x2_t __p0) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vqnegq_v((int8x16_t)__rev0, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vqneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vqneg_s64(int64x1_t __p0) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vqneg_v((int8x8_t)__p0, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqnegb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
-  return __ret;
-}
-#else
-__ai int8_t vqnegb_s8(int8_t __p0) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqnegb_s8(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqnegs_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
-  return __ret;
-}
-#else
-__ai int32_t vqnegs_s32(int32_t __p0) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqnegs_s32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
-  return __ret;
-}
-#else
-__ai int64_t vqnegd_s64(int64_t __p0) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqnegd_s64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqnegh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
-  return __ret;
-}
-#else
-__ai int16_t vqnegh_s16(int16_t __p0) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqnegh_s16(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqrdmulhs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrdmulhs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqrdmulhh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrdmulhh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_lane_s32(__p0_174, __p1_174, __p2_174) __extension__ ({ \
-  int32_t __s0_174 = __p0_174; \
-  int32x2_t __s1_174 = __p1_174; \
-  int32_t __ret_174; \
-  __ret_174 = vqrdmulhs_s32(__s0_174, vget_lane_s32(__s1_174, __p2_174)); \
-  __ret_174; \
-})
-#else
-#define vqrdmulhs_lane_s32(__p0_175, __p1_175, __p2_175) __extension__ ({ \
-  int32_t __s0_175 = __p0_175; \
-  int32x2_t __s1_175 = __p1_175; \
-  int32x2_t __rev1_175;  __rev1_175 = __builtin_shufflevector(__s1_175, __s1_175, 1, 0); \
-  int32_t __ret_175; \
-  __ret_175 = __noswap_vqrdmulhs_s32(__s0_175, __noswap_vget_lane_s32(__rev1_175, __p2_175)); \
-  __ret_175; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_lane_s16(__p0_176, __p1_176, __p2_176) __extension__ ({ \
-  int16_t __s0_176 = __p0_176; \
-  int16x4_t __s1_176 = __p1_176; \
-  int16_t __ret_176; \
-  __ret_176 = vqrdmulhh_s16(__s0_176, vget_lane_s16(__s1_176, __p2_176)); \
-  __ret_176; \
-})
-#else
-#define vqrdmulhh_lane_s16(__p0_177, __p1_177, __p2_177) __extension__ ({ \
-  int16_t __s0_177 = __p0_177; \
-  int16x4_t __s1_177 = __p1_177; \
-  int16x4_t __rev1_177;  __rev1_177 = __builtin_shufflevector(__s1_177, __s1_177, 3, 2, 1, 0); \
-  int16_t __ret_177; \
-  __ret_177 = __noswap_vqrdmulhh_s16(__s0_177, __noswap_vget_lane_s16(__rev1_177, __p2_177)); \
-  __ret_177; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhs_laneq_s32(__p0_178, __p1_178, __p2_178) __extension__ ({ \
-  int32_t __s0_178 = __p0_178; \
-  int32x4_t __s1_178 = __p1_178; \
-  int32_t __ret_178; \
-  __ret_178 = vqrdmulhs_s32(__s0_178, vgetq_lane_s32(__s1_178, __p2_178)); \
-  __ret_178; \
-})
-#else
-#define vqrdmulhs_laneq_s32(__p0_179, __p1_179, __p2_179) __extension__ ({ \
-  int32_t __s0_179 = __p0_179; \
-  int32x4_t __s1_179 = __p1_179; \
-  int32x4_t __rev1_179;  __rev1_179 = __builtin_shufflevector(__s1_179, __s1_179, 3, 2, 1, 0); \
-  int32_t __ret_179; \
-  __ret_179 = __noswap_vqrdmulhs_s32(__s0_179, __noswap_vgetq_lane_s32(__rev1_179, __p2_179)); \
-  __ret_179; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhh_laneq_s16(__p0_180, __p1_180, __p2_180) __extension__ ({ \
-  int16_t __s0_180 = __p0_180; \
-  int16x8_t __s1_180 = __p1_180; \
-  int16_t __ret_180; \
-  __ret_180 = vqrdmulhh_s16(__s0_180, vgetq_lane_s16(__s1_180, __p2_180)); \
-  __ret_180; \
-})
-#else
-#define vqrdmulhh_laneq_s16(__p0_181, __p1_181, __p2_181) __extension__ ({ \
-  int16_t __s0_181 = __p0_181; \
-  int16x8_t __s1_181 = __p1_181; \
-  int16x8_t __rev1_181;  __rev1_181 = __builtin_shufflevector(__s1_181, __s1_181, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_181; \
-  __ret_181 = __noswap_vqrdmulhh_s16(__s0_181, __noswap_vgetq_lane_s16(__rev1_181, __p2_181)); \
-  __ret_181; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __ret; \
-  __ret = vqrdmulhq_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulhq_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __noswap_vqrdmulhq_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __ret; \
-  __ret = vqrdmulhq_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulhq_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x8_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret; \
-  __ret = __noswap_vqrdmulhq_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __ret; \
-  __ret = vqrdmulh_s32(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulh_laneq_s32(__p0, __p1, __p2) __extension__ ({ \
-  int32x2_t __s0 = __p0; \
-  int32x4_t __s1 = __p1; \
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int32x2_t __ret; \
-  __ret = __noswap_vqrdmulh_s32(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __ret; \
-  __ret = vqrdmulh_s16(__s0, __builtin_shufflevector(__s1, __s1, __p2, __p2, __p2, __p2)); \
-  __ret; \
-})
-#else
-#define vqrdmulh_laneq_s16(__p0, __p1, __p2) __extension__ ({ \
-  int16x4_t __s0 = __p0; \
-  int16x8_t __s1 = __p1; \
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x4_t __ret; \
-  __ret = __noswap_vqrdmulh_s16(__rev0, __builtin_shufflevector(__rev1, __rev1, __p2, __p2, __p2, __p2)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqrshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqrshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqrshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqrshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqrshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqrshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqrshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqrshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqrshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqrshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqrshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqrshlh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u32(__p0_182, __p1_182, __p2_182) __extension__ ({ \
-  uint16x4_t __s0_182 = __p0_182; \
-  uint32x4_t __s1_182 = __p1_182; \
-  uint16x8_t __ret_182; \
-  __ret_182 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_182), (uint16x4_t)(vqrshrn_n_u32(__s1_182, __p2_182)))); \
-  __ret_182; \
-})
-#else
-#define vqrshrn_high_n_u32(__p0_183, __p1_183, __p2_183) __extension__ ({ \
-  uint16x4_t __s0_183 = __p0_183; \
-  uint32x4_t __s1_183 = __p1_183; \
-  uint16x4_t __rev0_183;  __rev0_183 = __builtin_shufflevector(__s0_183, __s0_183, 3, 2, 1, 0); \
-  uint32x4_t __rev1_183;  __rev1_183 = __builtin_shufflevector(__s1_183, __s1_183, 3, 2, 1, 0); \
-  uint16x8_t __ret_183; \
-  __ret_183 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_183), (uint16x4_t)(__noswap_vqrshrn_n_u32(__rev1_183, __p2_183)))); \
-  __ret_183 = __builtin_shufflevector(__ret_183, __ret_183, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_183; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u64(__p0_184, __p1_184, __p2_184) __extension__ ({ \
-  uint32x2_t __s0_184 = __p0_184; \
-  uint64x2_t __s1_184 = __p1_184; \
-  uint32x4_t __ret_184; \
-  __ret_184 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_184), (uint32x2_t)(vqrshrn_n_u64(__s1_184, __p2_184)))); \
-  __ret_184; \
-})
-#else
-#define vqrshrn_high_n_u64(__p0_185, __p1_185, __p2_185) __extension__ ({ \
-  uint32x2_t __s0_185 = __p0_185; \
-  uint64x2_t __s1_185 = __p1_185; \
-  uint32x2_t __rev0_185;  __rev0_185 = __builtin_shufflevector(__s0_185, __s0_185, 1, 0); \
-  uint64x2_t __rev1_185;  __rev1_185 = __builtin_shufflevector(__s1_185, __s1_185, 1, 0); \
-  uint32x4_t __ret_185; \
-  __ret_185 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_185), (uint32x2_t)(__noswap_vqrshrn_n_u64(__rev1_185, __p2_185)))); \
-  __ret_185 = __builtin_shufflevector(__ret_185, __ret_185, 3, 2, 1, 0); \
-  __ret_185; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_u16(__p0_186, __p1_186, __p2_186) __extension__ ({ \
-  uint8x8_t __s0_186 = __p0_186; \
-  uint16x8_t __s1_186 = __p1_186; \
-  uint8x16_t __ret_186; \
-  __ret_186 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_186), (uint8x8_t)(vqrshrn_n_u16(__s1_186, __p2_186)))); \
-  __ret_186; \
-})
-#else
-#define vqrshrn_high_n_u16(__p0_187, __p1_187, __p2_187) __extension__ ({ \
-  uint8x8_t __s0_187 = __p0_187; \
-  uint16x8_t __s1_187 = __p1_187; \
-  uint8x8_t __rev0_187;  __rev0_187 = __builtin_shufflevector(__s0_187, __s0_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_187;  __rev1_187 = __builtin_shufflevector(__s1_187, __s1_187, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_187; \
-  __ret_187 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_187), (uint8x8_t)(__noswap_vqrshrn_n_u16(__rev1_187, __p2_187)))); \
-  __ret_187 = __builtin_shufflevector(__ret_187, __ret_187, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_187; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s32(__p0_188, __p1_188, __p2_188) __extension__ ({ \
-  int16x4_t __s0_188 = __p0_188; \
-  int32x4_t __s1_188 = __p1_188; \
-  int16x8_t __ret_188; \
-  __ret_188 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_188), (int16x4_t)(vqrshrn_n_s32(__s1_188, __p2_188)))); \
-  __ret_188; \
-})
-#else
-#define vqrshrn_high_n_s32(__p0_189, __p1_189, __p2_189) __extension__ ({ \
-  int16x4_t __s0_189 = __p0_189; \
-  int32x4_t __s1_189 = __p1_189; \
-  int16x4_t __rev0_189;  __rev0_189 = __builtin_shufflevector(__s0_189, __s0_189, 3, 2, 1, 0); \
-  int32x4_t __rev1_189;  __rev1_189 = __builtin_shufflevector(__s1_189, __s1_189, 3, 2, 1, 0); \
-  int16x8_t __ret_189; \
-  __ret_189 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_189), (int16x4_t)(__noswap_vqrshrn_n_s32(__rev1_189, __p2_189)))); \
-  __ret_189 = __builtin_shufflevector(__ret_189, __ret_189, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_189; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s64(__p0_190, __p1_190, __p2_190) __extension__ ({ \
-  int32x2_t __s0_190 = __p0_190; \
-  int64x2_t __s1_190 = __p1_190; \
-  int32x4_t __ret_190; \
-  __ret_190 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_190), (int32x2_t)(vqrshrn_n_s64(__s1_190, __p2_190)))); \
-  __ret_190; \
-})
-#else
-#define vqrshrn_high_n_s64(__p0_191, __p1_191, __p2_191) __extension__ ({ \
-  int32x2_t __s0_191 = __p0_191; \
-  int64x2_t __s1_191 = __p1_191; \
-  int32x2_t __rev0_191;  __rev0_191 = __builtin_shufflevector(__s0_191, __s0_191, 1, 0); \
-  int64x2_t __rev1_191;  __rev1_191 = __builtin_shufflevector(__s1_191, __s1_191, 1, 0); \
-  int32x4_t __ret_191; \
-  __ret_191 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_191), (int32x2_t)(__noswap_vqrshrn_n_s64(__rev1_191, __p2_191)))); \
-  __ret_191 = __builtin_shufflevector(__ret_191, __ret_191, 3, 2, 1, 0); \
-  __ret_191; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrn_high_n_s16(__p0_192, __p1_192, __p2_192) __extension__ ({ \
-  int8x8_t __s0_192 = __p0_192; \
-  int16x8_t __s1_192 = __p1_192; \
-  int8x16_t __ret_192; \
-  __ret_192 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_192), (int8x8_t)(vqrshrn_n_s16(__s1_192, __p2_192)))); \
-  __ret_192; \
-})
-#else
-#define vqrshrn_high_n_s16(__p0_193, __p1_193, __p2_193) __extension__ ({ \
-  int8x8_t __s0_193 = __p0_193; \
-  int16x8_t __s1_193 = __p1_193; \
-  int8x8_t __rev0_193;  __rev0_193 = __builtin_shufflevector(__s0_193, __s0_193, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_193;  __rev1_193 = __builtin_shufflevector(__s1_193, __s1_193, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_193; \
-  __ret_193 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_193), (int8x8_t)(__noswap_vqrshrn_n_s16(__rev1_193, __p2_193)))); \
-  __ret_193 = __builtin_shufflevector(__ret_193, __ret_193, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_193; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqrshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqrshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqrshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s32(__p0_194, __p1_194, __p2_194) __extension__ ({ \
-  int16x4_t __s0_194 = __p0_194; \
-  int32x4_t __s1_194 = __p1_194; \
-  int16x8_t __ret_194; \
-  __ret_194 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_194), (int16x4_t)(vqrshrun_n_s32(__s1_194, __p2_194)))); \
-  __ret_194; \
-})
-#else
-#define vqrshrun_high_n_s32(__p0_195, __p1_195, __p2_195) __extension__ ({ \
-  int16x4_t __s0_195 = __p0_195; \
-  int32x4_t __s1_195 = __p1_195; \
-  int16x4_t __rev0_195;  __rev0_195 = __builtin_shufflevector(__s0_195, __s0_195, 3, 2, 1, 0); \
-  int32x4_t __rev1_195;  __rev1_195 = __builtin_shufflevector(__s1_195, __s1_195, 3, 2, 1, 0); \
-  int16x8_t __ret_195; \
-  __ret_195 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_195), (int16x4_t)(__noswap_vqrshrun_n_s32(__rev1_195, __p2_195)))); \
-  __ret_195 = __builtin_shufflevector(__ret_195, __ret_195, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_195; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s64(__p0_196, __p1_196, __p2_196) __extension__ ({ \
-  int32x2_t __s0_196 = __p0_196; \
-  int64x2_t __s1_196 = __p1_196; \
-  int32x4_t __ret_196; \
-  __ret_196 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_196), (int32x2_t)(vqrshrun_n_s64(__s1_196, __p2_196)))); \
-  __ret_196; \
-})
-#else
-#define vqrshrun_high_n_s64(__p0_197, __p1_197, __p2_197) __extension__ ({ \
-  int32x2_t __s0_197 = __p0_197; \
-  int64x2_t __s1_197 = __p1_197; \
-  int32x2_t __rev0_197;  __rev0_197 = __builtin_shufflevector(__s0_197, __s0_197, 1, 0); \
-  int64x2_t __rev1_197;  __rev1_197 = __builtin_shufflevector(__s1_197, __s1_197, 1, 0); \
-  int32x4_t __ret_197; \
-  __ret_197 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_197), (int32x2_t)(__noswap_vqrshrun_n_s64(__rev1_197, __p2_197)))); \
-  __ret_197 = __builtin_shufflevector(__ret_197, __ret_197, 3, 2, 1, 0); \
-  __ret_197; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrun_high_n_s16(__p0_198, __p1_198, __p2_198) __extension__ ({ \
-  int8x8_t __s0_198 = __p0_198; \
-  int16x8_t __s1_198 = __p1_198; \
-  int8x16_t __ret_198; \
-  __ret_198 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_198), (int8x8_t)(vqrshrun_n_s16(__s1_198, __p2_198)))); \
-  __ret_198; \
-})
-#else
-#define vqrshrun_high_n_s16(__p0_199, __p1_199, __p2_199) __extension__ ({ \
-  int8x8_t __s0_199 = __p0_199; \
-  int16x8_t __s1_199 = __p1_199; \
-  int8x8_t __rev0_199;  __rev0_199 = __builtin_shufflevector(__s0_199, __s0_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_199;  __rev1_199 = __builtin_shufflevector(__s1_199, __s1_199, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_199; \
-  __ret_199 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_199), (int8x8_t)(__noswap_vqrshrun_n_s16(__rev1_199, __p2_199)))); \
-  __ret_199 = __builtin_shufflevector(__ret_199, __ret_199, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_199; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqrshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqrshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqrshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqrshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqshlb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqshlb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqshls_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqshls_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqshlh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqshlh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqshlb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqshlb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqshls_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqshls_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqshlh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqshlh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_u8(__p0, __p1) __extension__ ({ \
-  uint8_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshlb_n_u8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshls_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vqshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshlh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlb_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlb_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshls_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshls_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshlh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlub_n_s8(__p0, __p1) __extension__ ({ \
-  int8_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshlub_n_s8(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlus_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshlus_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshlud_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vqshlud_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshluh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshluh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u32(__p0_200, __p1_200, __p2_200) __extension__ ({ \
-  uint16x4_t __s0_200 = __p0_200; \
-  uint32x4_t __s1_200 = __p1_200; \
-  uint16x8_t __ret_200; \
-  __ret_200 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_200), (uint16x4_t)(vqshrn_n_u32(__s1_200, __p2_200)))); \
-  __ret_200; \
-})
-#else
-#define vqshrn_high_n_u32(__p0_201, __p1_201, __p2_201) __extension__ ({ \
-  uint16x4_t __s0_201 = __p0_201; \
-  uint32x4_t __s1_201 = __p1_201; \
-  uint16x4_t __rev0_201;  __rev0_201 = __builtin_shufflevector(__s0_201, __s0_201, 3, 2, 1, 0); \
-  uint32x4_t __rev1_201;  __rev1_201 = __builtin_shufflevector(__s1_201, __s1_201, 3, 2, 1, 0); \
-  uint16x8_t __ret_201; \
-  __ret_201 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_201), (uint16x4_t)(__noswap_vqshrn_n_u32(__rev1_201, __p2_201)))); \
-  __ret_201 = __builtin_shufflevector(__ret_201, __ret_201, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_201; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u64(__p0_202, __p1_202, __p2_202) __extension__ ({ \
-  uint32x2_t __s0_202 = __p0_202; \
-  uint64x2_t __s1_202 = __p1_202; \
-  uint32x4_t __ret_202; \
-  __ret_202 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_202), (uint32x2_t)(vqshrn_n_u64(__s1_202, __p2_202)))); \
-  __ret_202; \
-})
-#else
-#define vqshrn_high_n_u64(__p0_203, __p1_203, __p2_203) __extension__ ({ \
-  uint32x2_t __s0_203 = __p0_203; \
-  uint64x2_t __s1_203 = __p1_203; \
-  uint32x2_t __rev0_203;  __rev0_203 = __builtin_shufflevector(__s0_203, __s0_203, 1, 0); \
-  uint64x2_t __rev1_203;  __rev1_203 = __builtin_shufflevector(__s1_203, __s1_203, 1, 0); \
-  uint32x4_t __ret_203; \
-  __ret_203 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_203), (uint32x2_t)(__noswap_vqshrn_n_u64(__rev1_203, __p2_203)))); \
-  __ret_203 = __builtin_shufflevector(__ret_203, __ret_203, 3, 2, 1, 0); \
-  __ret_203; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_u16(__p0_204, __p1_204, __p2_204) __extension__ ({ \
-  uint8x8_t __s0_204 = __p0_204; \
-  uint16x8_t __s1_204 = __p1_204; \
-  uint8x16_t __ret_204; \
-  __ret_204 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_204), (uint8x8_t)(vqshrn_n_u16(__s1_204, __p2_204)))); \
-  __ret_204; \
-})
-#else
-#define vqshrn_high_n_u16(__p0_205, __p1_205, __p2_205) __extension__ ({ \
-  uint8x8_t __s0_205 = __p0_205; \
-  uint16x8_t __s1_205 = __p1_205; \
-  uint8x8_t __rev0_205;  __rev0_205 = __builtin_shufflevector(__s0_205, __s0_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_205;  __rev1_205 = __builtin_shufflevector(__s1_205, __s1_205, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_205; \
-  __ret_205 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_205), (uint8x8_t)(__noswap_vqshrn_n_u16(__rev1_205, __p2_205)))); \
-  __ret_205 = __builtin_shufflevector(__ret_205, __ret_205, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_205; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s32(__p0_206, __p1_206, __p2_206) __extension__ ({ \
-  int16x4_t __s0_206 = __p0_206; \
-  int32x4_t __s1_206 = __p1_206; \
-  int16x8_t __ret_206; \
-  __ret_206 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_206), (int16x4_t)(vqshrn_n_s32(__s1_206, __p2_206)))); \
-  __ret_206; \
-})
-#else
-#define vqshrn_high_n_s32(__p0_207, __p1_207, __p2_207) __extension__ ({ \
-  int16x4_t __s0_207 = __p0_207; \
-  int32x4_t __s1_207 = __p1_207; \
-  int16x4_t __rev0_207;  __rev0_207 = __builtin_shufflevector(__s0_207, __s0_207, 3, 2, 1, 0); \
-  int32x4_t __rev1_207;  __rev1_207 = __builtin_shufflevector(__s1_207, __s1_207, 3, 2, 1, 0); \
-  int16x8_t __ret_207; \
-  __ret_207 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_207), (int16x4_t)(__noswap_vqshrn_n_s32(__rev1_207, __p2_207)))); \
-  __ret_207 = __builtin_shufflevector(__ret_207, __ret_207, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_207; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s64(__p0_208, __p1_208, __p2_208) __extension__ ({ \
-  int32x2_t __s0_208 = __p0_208; \
-  int64x2_t __s1_208 = __p1_208; \
-  int32x4_t __ret_208; \
-  __ret_208 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_208), (int32x2_t)(vqshrn_n_s64(__s1_208, __p2_208)))); \
-  __ret_208; \
-})
-#else
-#define vqshrn_high_n_s64(__p0_209, __p1_209, __p2_209) __extension__ ({ \
-  int32x2_t __s0_209 = __p0_209; \
-  int64x2_t __s1_209 = __p1_209; \
-  int32x2_t __rev0_209;  __rev0_209 = __builtin_shufflevector(__s0_209, __s0_209, 1, 0); \
-  int64x2_t __rev1_209;  __rev1_209 = __builtin_shufflevector(__s1_209, __s1_209, 1, 0); \
-  int32x4_t __ret_209; \
-  __ret_209 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_209), (int32x2_t)(__noswap_vqshrn_n_s64(__rev1_209, __p2_209)))); \
-  __ret_209 = __builtin_shufflevector(__ret_209, __ret_209, 3, 2, 1, 0); \
-  __ret_209; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrn_high_n_s16(__p0_210, __p1_210, __p2_210) __extension__ ({ \
-  int8x8_t __s0_210 = __p0_210; \
-  int16x8_t __s1_210 = __p1_210; \
-  int8x16_t __ret_210; \
-  __ret_210 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_210), (int8x8_t)(vqshrn_n_s16(__s1_210, __p2_210)))); \
-  __ret_210; \
-})
-#else
-#define vqshrn_high_n_s16(__p0_211, __p1_211, __p2_211) __extension__ ({ \
-  int8x8_t __s0_211 = __p0_211; \
-  int16x8_t __s1_211 = __p1_211; \
-  int8x8_t __rev0_211;  __rev0_211 = __builtin_shufflevector(__s0_211, __s0_211, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_211;  __rev1_211 = __builtin_shufflevector(__s1_211, __s1_211, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_211; \
-  __ret_211 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_211), (int8x8_t)(__noswap_vqshrn_n_s16(__rev1_211, __p2_211)))); \
-  __ret_211 = __builtin_shufflevector(__ret_211, __ret_211, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_211; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_u32(__p0, __p1) __extension__ ({ \
-  uint32_t __s0 = __p0; \
-  uint16_t __ret; \
-  __ret = (uint16_t) __builtin_neon_vqshrns_n_u32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint32_t __ret; \
-  __ret = (uint32_t) __builtin_neon_vqshrnd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_u16(__p0, __p1) __extension__ ({ \
-  uint16_t __s0 = __p0; \
-  uint8_t __ret; \
-  __ret = (uint8_t) __builtin_neon_vqshrnh_n_u16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshrns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrnd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrnh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrnh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s32(__p0_212, __p1_212, __p2_212) __extension__ ({ \
-  int16x4_t __s0_212 = __p0_212; \
-  int32x4_t __s1_212 = __p1_212; \
-  int16x8_t __ret_212; \
-  __ret_212 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_212), (int16x4_t)(vqshrun_n_s32(__s1_212, __p2_212)))); \
-  __ret_212; \
-})
-#else
-#define vqshrun_high_n_s32(__p0_213, __p1_213, __p2_213) __extension__ ({ \
-  int16x4_t __s0_213 = __p0_213; \
-  int32x4_t __s1_213 = __p1_213; \
-  int16x4_t __rev0_213;  __rev0_213 = __builtin_shufflevector(__s0_213, __s0_213, 3, 2, 1, 0); \
-  int32x4_t __rev1_213;  __rev1_213 = __builtin_shufflevector(__s1_213, __s1_213, 3, 2, 1, 0); \
-  int16x8_t __ret_213; \
-  __ret_213 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_213), (int16x4_t)(__noswap_vqshrun_n_s32(__rev1_213, __p2_213)))); \
-  __ret_213 = __builtin_shufflevector(__ret_213, __ret_213, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_213; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s64(__p0_214, __p1_214, __p2_214) __extension__ ({ \
-  int32x2_t __s0_214 = __p0_214; \
-  int64x2_t __s1_214 = __p1_214; \
-  int32x4_t __ret_214; \
-  __ret_214 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_214), (int32x2_t)(vqshrun_n_s64(__s1_214, __p2_214)))); \
-  __ret_214; \
-})
-#else
-#define vqshrun_high_n_s64(__p0_215, __p1_215, __p2_215) __extension__ ({ \
-  int32x2_t __s0_215 = __p0_215; \
-  int64x2_t __s1_215 = __p1_215; \
-  int32x2_t __rev0_215;  __rev0_215 = __builtin_shufflevector(__s0_215, __s0_215, 1, 0); \
-  int64x2_t __rev1_215;  __rev1_215 = __builtin_shufflevector(__s1_215, __s1_215, 1, 0); \
-  int32x4_t __ret_215; \
-  __ret_215 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_215), (int32x2_t)(__noswap_vqshrun_n_s64(__rev1_215, __p2_215)))); \
-  __ret_215 = __builtin_shufflevector(__ret_215, __ret_215, 3, 2, 1, 0); \
-  __ret_215; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrun_high_n_s16(__p0_216, __p1_216, __p2_216) __extension__ ({ \
-  int8x8_t __s0_216 = __p0_216; \
-  int16x8_t __s1_216 = __p1_216; \
-  int8x16_t __ret_216; \
-  __ret_216 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_216), (int8x8_t)(vqshrun_n_s16(__s1_216, __p2_216)))); \
-  __ret_216; \
-})
-#else
-#define vqshrun_high_n_s16(__p0_217, __p1_217, __p2_217) __extension__ ({ \
-  int8x8_t __s0_217 = __p0_217; \
-  int16x8_t __s1_217 = __p1_217; \
-  int8x8_t __rev0_217;  __rev0_217 = __builtin_shufflevector(__s0_217, __s0_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_217;  __rev1_217 = __builtin_shufflevector(__s1_217, __s1_217, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_217; \
-  __ret_217 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_217), (int8x8_t)(__noswap_vqshrun_n_s16(__rev1_217, __p2_217)))); \
-  __ret_217 = __builtin_shufflevector(__ret_217, __ret_217, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_217; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshruns_n_s32(__p0, __p1) __extension__ ({ \
-  int32_t __s0 = __p0; \
-  int16_t __ret; \
-  __ret = (int16_t) __builtin_neon_vqshruns_n_s32(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrund_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int32_t __ret; \
-  __ret = (int32_t) __builtin_neon_vqshrund_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vqshrunh_n_s16(__p0, __p1) __extension__ ({ \
-  int16_t __s0 = __p0; \
-  int8_t __ret; \
-  __ret = (int8_t) __builtin_neon_vqshrunh_n_s16(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vqsubb_u8(uint8_t __p0, uint8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vqsubb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vqsubs_u32(uint32_t __p0, uint32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vqsubs_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vqsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vqsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vqsubh_u16(uint16_t __p0, uint16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vqsubh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vqsubb_s8(int8_t __p0, int8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vqsubb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-__ai int32_t __noswap_vqsubs_s32(int32_t __p0, int32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vqsubs_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vqsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vqsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-__ai int16_t __noswap_vqsubh_s16(int16_t __p0, int16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vqsubh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbl1_p8(poly8x16_t __p0, uint8x8_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbl1q_p8(poly8x16_t __p0, uint8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbl1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbl1q_s8(int8x16_t __p0, uint8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbl1q_s8(int8x16_t __p0, uint8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbl1_u8(uint8x16_t __p0, uint8x8_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbl1_s8(int8x16_t __p0, uint8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbl1_s8(int8x16_t __p0, uint8x8_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl1_v((int8x16_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbl2_p8(poly8x16x2_t __p0, uint8x8_t __p1) {
-  poly8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbl2q_p8(poly8x16x2_t __p0, uint8x16_t __p1) {
-  poly8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbl2q_u8(uint8x16x2_t __p0, uint8x16_t __p1) {
-  uint8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, uint8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbl2q_s8(int8x16x2_t __p0, uint8x16_t __p1) {
-  int8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl2q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbl2_u8(uint8x16x2_t __p0, uint8x8_t __p1) {
-  uint8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, uint8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbl2_s8(int8x16x2_t __p0, uint8x8_t __p1) {
-  int8x16x2_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl2_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbl3_p8(poly8x16x3_t __p0, uint8x8_t __p1) {
-  poly8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbl3q_p8(poly8x16x3_t __p0, uint8x16_t __p1) {
-  poly8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbl3q_u8(uint8x16x3_t __p0, uint8x16_t __p1) {
-  uint8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, uint8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbl3q_s8(int8x16x3_t __p0, uint8x16_t __p1) {
-  int8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl3q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbl3_u8(uint8x16x3_t __p0, uint8x8_t __p1) {
-  uint8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, uint8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbl3_s8(int8x16x3_t __p0, uint8x8_t __p1) {
-  int8x16x3_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl3_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbl4_p8(poly8x16x4_t __p0, uint8x8_t __p1) {
-  poly8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbl4q_p8(poly8x16x4_t __p0, uint8x16_t __p1) {
-  poly8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbl4q_u8(uint8x16x4_t __p0, uint8x16_t __p1) {
-  uint8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, uint8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbl4q_s8(int8x16x4_t __p0, uint8x16_t __p1) {
-  int8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbl4q_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbl4_u8(uint8x16x4_t __p0, uint8x8_t __p1) {
-  uint8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, uint8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__p0.val[0], (int8x16_t)__p0.val[1], (int8x16_t)__p0.val[2], (int8x16_t)__p0.val[3], (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbl4_s8(int8x16x4_t __p0, uint8x8_t __p1) {
-  int8x16x4_t __rev0;
-  __rev0.val[0] = __builtin_shufflevector(__p0.val[0], __p0.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[1] = __builtin_shufflevector(__p0.val[1], __p0.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[2] = __builtin_shufflevector(__p0.val[2], __p0.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev0.val[3] = __builtin_shufflevector(__p0.val[3], __p0.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbl4_v((int8x16_t)__rev0.val[0], (int8x16_t)__rev0.val[1], (int8x16_t)__rev0.val[2], (int8x16_t)__rev0.val[3], (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbx1_p8(poly8x8_t __p0, poly8x16_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbx1q_p8(poly8x16_t __p0, poly8x16_t __p1, uint8x16_t __p2) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbx1q_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, uint8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__p0, (int8x16_t)__p1, (int8x16_t)__p2, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbx1q_s8(int8x16_t __p0, int8x16_t __p1, uint8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx1q_v((int8x16_t)__rev0, (int8x16_t)__rev1, (int8x16_t)__rev2, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbx1_u8(uint8x8_t __p0, uint8x16_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, uint8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__p0, (int8x16_t)__p1, (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbx1_s8(int8x8_t __p0, int8x16_t __p1, uint8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx1_v((int8x8_t)__rev0, (int8x16_t)__rev1, (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbx2_p8(poly8x8_t __p0, poly8x16x2_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbx2q_p8(poly8x16_t __p0, poly8x16x2_t __p1, uint8x16_t __p2) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbx2q_u8(uint8x16_t __p0, uint8x16x2_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, uint8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p2, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbx2q_s8(int8x16_t __p0, int8x16x2_t __p1, uint8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx2q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev2, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbx2_u8(uint8x8_t __p0, uint8x16x2_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, uint8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbx2_s8(int8x8_t __p0, int8x16x2_t __p1, uint8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x2_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx2_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbx3_p8(poly8x8_t __p0, poly8x16x3_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbx3q_p8(poly8x16_t __p0, poly8x16x3_t __p1, uint8x16_t __p2) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbx3q_u8(uint8x16_t __p0, uint8x16x3_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, uint8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p2, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbx3q_s8(int8x16_t __p0, int8x16x3_t __p1, uint8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx3q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev2, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbx3_u8(uint8x8_t __p0, uint8x16x3_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, uint8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbx3_s8(int8x8_t __p0, int8x16x3_t __p1, uint8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x3_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx3_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vqtbx4_p8(poly8x8_t __p0, poly8x16x4_t __p1, uint8x8_t __p2) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vqtbx4q_p8(poly8x16_t __p0, poly8x16x4_t __p1, uint8x16_t __p2) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vqtbx4q_u8(uint8x16_t __p0, uint8x16x4_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, uint8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x16_t)__p2, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vqtbx4q_s8(int8x16_t __p0, int8x16x4_t __p1, uint8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vqtbx4q_v((int8x16_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x16_t)__rev2, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vqtbx4_u8(uint8x8_t __p0, uint8x16x4_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, uint8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__p0, (int8x16_t)__p1.val[0], (int8x16_t)__p1.val[1], (int8x16_t)__p1.val[2], (int8x16_t)__p1.val[3], (int8x8_t)__p2, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vqtbx4_s8(int8x8_t __p0, int8x16x4_t __p1, uint8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16x4_t __rev1;
-  __rev1.val[0] = __builtin_shufflevector(__p1.val[0], __p1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[1] = __builtin_shufflevector(__p1.val[1], __p1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[2] = __builtin_shufflevector(__p1.val[2], __p1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  __rev1.val[3] = __builtin_shufflevector(__p1.val[3], __p1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vqtbx4_v((int8x8_t)__rev0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], (int8x8_t)__rev2, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vraddhn_u32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vraddhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vraddhn_u32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vraddhn_u64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vraddhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vraddhn_u64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vraddhn_u16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint8x16_t vraddhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vraddhn_u16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vraddhn_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vraddhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vraddhn_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vraddhn_s64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vraddhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vraddhn_s64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vraddhn_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int8x16_t vraddhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vraddhn_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 4);
-  return __ret;
-}
-#else
-__ai poly8x8_t vrbit_p8(poly8x8_t __p0) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = (poly8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 4);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 36);
-  return __ret;
-}
-#else
-__ai poly8x16_t vrbitq_p8(poly8x16_t __p0) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = (poly8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 36);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vrbitq_u8(uint8x16_t __p0) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrbitq_s8(int8x16_t __p0) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__p0, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vrbitq_s8(int8x16_t __p0) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vrbitq_v((int8x16_t)__rev0, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vrbit_u8(uint8x8_t __p0) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vrbit_s8(int8x8_t __p0) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__p0, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vrbit_s8(int8x8_t __p0) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vrbit_v((int8x8_t)__rev0, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrecpeq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrecpeq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrecpe_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecpe_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vrecped_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
-  return __ret;
-}
-#else
-__ai float64_t vrecped_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecped_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrecpes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vrecpes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpes_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrecpsq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrecpsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrecps_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrecps_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float64_t vrecpsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float32_t vrecpss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vrecpxd_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
-  return __ret;
-}
-#else
-__ai float64_t vrecpxd_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrecpxd_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrecpxs_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vrecpxs_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrecpxs_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vrshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vrshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vrshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vrshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vrshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vrshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u32(__p0_218, __p1_218, __p2_218) __extension__ ({ \
-  uint16x4_t __s0_218 = __p0_218; \
-  uint32x4_t __s1_218 = __p1_218; \
-  uint16x8_t __ret_218; \
-  __ret_218 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_218), (uint16x4_t)(vrshrn_n_u32(__s1_218, __p2_218)))); \
-  __ret_218; \
-})
-#else
-#define vrshrn_high_n_u32(__p0_219, __p1_219, __p2_219) __extension__ ({ \
-  uint16x4_t __s0_219 = __p0_219; \
-  uint32x4_t __s1_219 = __p1_219; \
-  uint16x4_t __rev0_219;  __rev0_219 = __builtin_shufflevector(__s0_219, __s0_219, 3, 2, 1, 0); \
-  uint32x4_t __rev1_219;  __rev1_219 = __builtin_shufflevector(__s1_219, __s1_219, 3, 2, 1, 0); \
-  uint16x8_t __ret_219; \
-  __ret_219 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_219), (uint16x4_t)(__noswap_vrshrn_n_u32(__rev1_219, __p2_219)))); \
-  __ret_219 = __builtin_shufflevector(__ret_219, __ret_219, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_219; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u64(__p0_220, __p1_220, __p2_220) __extension__ ({ \
-  uint32x2_t __s0_220 = __p0_220; \
-  uint64x2_t __s1_220 = __p1_220; \
-  uint32x4_t __ret_220; \
-  __ret_220 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_220), (uint32x2_t)(vrshrn_n_u64(__s1_220, __p2_220)))); \
-  __ret_220; \
-})
-#else
-#define vrshrn_high_n_u64(__p0_221, __p1_221, __p2_221) __extension__ ({ \
-  uint32x2_t __s0_221 = __p0_221; \
-  uint64x2_t __s1_221 = __p1_221; \
-  uint32x2_t __rev0_221;  __rev0_221 = __builtin_shufflevector(__s0_221, __s0_221, 1, 0); \
-  uint64x2_t __rev1_221;  __rev1_221 = __builtin_shufflevector(__s1_221, __s1_221, 1, 0); \
-  uint32x4_t __ret_221; \
-  __ret_221 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_221), (uint32x2_t)(__noswap_vrshrn_n_u64(__rev1_221, __p2_221)))); \
-  __ret_221 = __builtin_shufflevector(__ret_221, __ret_221, 3, 2, 1, 0); \
-  __ret_221; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_u16(__p0_222, __p1_222, __p2_222) __extension__ ({ \
-  uint8x8_t __s0_222 = __p0_222; \
-  uint16x8_t __s1_222 = __p1_222; \
-  uint8x16_t __ret_222; \
-  __ret_222 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_222), (uint8x8_t)(vrshrn_n_u16(__s1_222, __p2_222)))); \
-  __ret_222; \
-})
-#else
-#define vrshrn_high_n_u16(__p0_223, __p1_223, __p2_223) __extension__ ({ \
-  uint8x8_t __s0_223 = __p0_223; \
-  uint16x8_t __s1_223 = __p1_223; \
-  uint8x8_t __rev0_223;  __rev0_223 = __builtin_shufflevector(__s0_223, __s0_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_223;  __rev1_223 = __builtin_shufflevector(__s1_223, __s1_223, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_223; \
-  __ret_223 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_223), (uint8x8_t)(__noswap_vrshrn_n_u16(__rev1_223, __p2_223)))); \
-  __ret_223 = __builtin_shufflevector(__ret_223, __ret_223, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_223; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s32(__p0_224, __p1_224, __p2_224) __extension__ ({ \
-  int16x4_t __s0_224 = __p0_224; \
-  int32x4_t __s1_224 = __p1_224; \
-  int16x8_t __ret_224; \
-  __ret_224 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_224), (int16x4_t)(vrshrn_n_s32(__s1_224, __p2_224)))); \
-  __ret_224; \
-})
-#else
-#define vrshrn_high_n_s32(__p0_225, __p1_225, __p2_225) __extension__ ({ \
-  int16x4_t __s0_225 = __p0_225; \
-  int32x4_t __s1_225 = __p1_225; \
-  int16x4_t __rev0_225;  __rev0_225 = __builtin_shufflevector(__s0_225, __s0_225, 3, 2, 1, 0); \
-  int32x4_t __rev1_225;  __rev1_225 = __builtin_shufflevector(__s1_225, __s1_225, 3, 2, 1, 0); \
-  int16x8_t __ret_225; \
-  __ret_225 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_225), (int16x4_t)(__noswap_vrshrn_n_s32(__rev1_225, __p2_225)))); \
-  __ret_225 = __builtin_shufflevector(__ret_225, __ret_225, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_225; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s64(__p0_226, __p1_226, __p2_226) __extension__ ({ \
-  int32x2_t __s0_226 = __p0_226; \
-  int64x2_t __s1_226 = __p1_226; \
-  int32x4_t __ret_226; \
-  __ret_226 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_226), (int32x2_t)(vrshrn_n_s64(__s1_226, __p2_226)))); \
-  __ret_226; \
-})
-#else
-#define vrshrn_high_n_s64(__p0_227, __p1_227, __p2_227) __extension__ ({ \
-  int32x2_t __s0_227 = __p0_227; \
-  int64x2_t __s1_227 = __p1_227; \
-  int32x2_t __rev0_227;  __rev0_227 = __builtin_shufflevector(__s0_227, __s0_227, 1, 0); \
-  int64x2_t __rev1_227;  __rev1_227 = __builtin_shufflevector(__s1_227, __s1_227, 1, 0); \
-  int32x4_t __ret_227; \
-  __ret_227 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_227), (int32x2_t)(__noswap_vrshrn_n_s64(__rev1_227, __p2_227)))); \
-  __ret_227 = __builtin_shufflevector(__ret_227, __ret_227, 3, 2, 1, 0); \
-  __ret_227; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrshrn_high_n_s16(__p0_228, __p1_228, __p2_228) __extension__ ({ \
-  int8x8_t __s0_228 = __p0_228; \
-  int16x8_t __s1_228 = __p1_228; \
-  int8x16_t __ret_228; \
-  __ret_228 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_228), (int8x8_t)(vrshrn_n_s16(__s1_228, __p2_228)))); \
-  __ret_228; \
-})
-#else
-#define vrshrn_high_n_s16(__p0_229, __p1_229, __p2_229) __extension__ ({ \
-  int8x8_t __s0_229 = __p0_229; \
-  int16x8_t __s1_229 = __p1_229; \
-  int8x8_t __rev0_229;  __rev0_229 = __builtin_shufflevector(__s0_229, __s0_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_229;  __rev1_229 = __builtin_shufflevector(__s1_229, __s1_229, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_229; \
-  __ret_229 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_229), (int8x8_t)(__noswap_vrshrn_n_s16(__rev1_229, __p2_229)))); \
-  __ret_229 = __builtin_shufflevector(__ret_229, __ret_229, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_229; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrsqrteq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrsqrteq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrsqrte_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrte_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vrsqrted_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
-  return __ret;
-}
-#else
-__ai float64_t vrsqrted_f64(float64_t __p0) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrted_f64(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrsqrtes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
-  return __ret;
-}
-#else
-__ai float32_t vrsqrtes_f32(float32_t __p0) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtes_f32(__p0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__p0, (int8x16_t)__p1, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vrsqrtsq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vrsqrtsq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vrsqrts_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vrsqrts_v((int8x8_t)__p0, (int8x8_t)__p1, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float64_t vrsqrtsd_f64(float64_t __p0, float64_t __p1) {
-  float64_t __ret;
-  __ret = (float64_t) __builtin_neon_vrsqrtsd_f64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai float32_t vrsqrtss_f32(float32_t __p0, float32_t __p1) {
-  float32_t __ret;
-  __ret = (float32_t) __builtin_neon_vrsqrtss_f32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vrsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vrsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vrsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vrsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vrsubhn_u32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vrsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vrsubhn_u32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vrsubhn_u64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vrsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vrsubhn_u64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vrsubhn_u16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint8x16_t vrsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vrsubhn_u16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vrsubhn_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vrsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vrsubhn_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vrsubhn_s64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vrsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vrsubhn_s64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vrsubhn_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int8x16_t vrsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vrsubhn_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vset_lane_i64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsetq_lane_i64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__rev1, __p2); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#define __noswap_vsetq_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __ret; \
-  __ret = (float64x2_t) __builtin_neon_vsetq_lane_f64(__s0, (int8x16_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#else
-#define vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#define __noswap_vset_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64_t __s0 = __p0; \
-  float64x1_t __s1 = __p1; \
-  float64x1_t __ret; \
-  __ret = (float64x1_t) __builtin_neon_vset_lane_f64(__s0, (int8x8_t)__s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vshld_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vshld_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vshld_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vshld_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshld_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshld_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshld_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshld_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u8(__p0_230, __p1_230) __extension__ ({ \
-  uint8x16_t __s0_230 = __p0_230; \
-  uint16x8_t __ret_230; \
-  __ret_230 = (uint16x8_t)(vshll_n_u8(vget_high_u8(__s0_230), __p1_230)); \
-  __ret_230; \
-})
-#else
-#define vshll_high_n_u8(__p0_231, __p1_231) __extension__ ({ \
-  uint8x16_t __s0_231 = __p0_231; \
-  uint8x16_t __rev0_231;  __rev0_231 = __builtin_shufflevector(__s0_231, __s0_231, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __ret_231; \
-  __ret_231 = (uint16x8_t)(__noswap_vshll_n_u8(__noswap_vget_high_u8(__rev0_231), __p1_231)); \
-  __ret_231 = __builtin_shufflevector(__ret_231, __ret_231, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_231; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u32(__p0_232, __p1_232) __extension__ ({ \
-  uint32x4_t __s0_232 = __p0_232; \
-  uint64x2_t __ret_232; \
-  __ret_232 = (uint64x2_t)(vshll_n_u32(vget_high_u32(__s0_232), __p1_232)); \
-  __ret_232; \
-})
-#else
-#define vshll_high_n_u32(__p0_233, __p1_233) __extension__ ({ \
-  uint32x4_t __s0_233 = __p0_233; \
-  uint32x4_t __rev0_233;  __rev0_233 = __builtin_shufflevector(__s0_233, __s0_233, 3, 2, 1, 0); \
-  uint64x2_t __ret_233; \
-  __ret_233 = (uint64x2_t)(__noswap_vshll_n_u32(__noswap_vget_high_u32(__rev0_233), __p1_233)); \
-  __ret_233 = __builtin_shufflevector(__ret_233, __ret_233, 1, 0); \
-  __ret_233; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_u16(__p0_234, __p1_234) __extension__ ({ \
-  uint16x8_t __s0_234 = __p0_234; \
-  uint32x4_t __ret_234; \
-  __ret_234 = (uint32x4_t)(vshll_n_u16(vget_high_u16(__s0_234), __p1_234)); \
-  __ret_234; \
-})
-#else
-#define vshll_high_n_u16(__p0_235, __p1_235) __extension__ ({ \
-  uint16x8_t __s0_235 = __p0_235; \
-  uint16x8_t __rev0_235;  __rev0_235 = __builtin_shufflevector(__s0_235, __s0_235, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint32x4_t __ret_235; \
-  __ret_235 = (uint32x4_t)(__noswap_vshll_n_u16(__noswap_vget_high_u16(__rev0_235), __p1_235)); \
-  __ret_235 = __builtin_shufflevector(__ret_235, __ret_235, 3, 2, 1, 0); \
-  __ret_235; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s8(__p0_236, __p1_236) __extension__ ({ \
-  int8x16_t __s0_236 = __p0_236; \
-  int16x8_t __ret_236; \
-  __ret_236 = (int16x8_t)(vshll_n_s8(vget_high_s8(__s0_236), __p1_236)); \
-  __ret_236; \
-})
-#else
-#define vshll_high_n_s8(__p0_237, __p1_237) __extension__ ({ \
-  int8x16_t __s0_237 = __p0_237; \
-  int8x16_t __rev0_237;  __rev0_237 = __builtin_shufflevector(__s0_237, __s0_237, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __ret_237; \
-  __ret_237 = (int16x8_t)(__noswap_vshll_n_s8(__noswap_vget_high_s8(__rev0_237), __p1_237)); \
-  __ret_237 = __builtin_shufflevector(__ret_237, __ret_237, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_237; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s32(__p0_238, __p1_238) __extension__ ({ \
-  int32x4_t __s0_238 = __p0_238; \
-  int64x2_t __ret_238; \
-  __ret_238 = (int64x2_t)(vshll_n_s32(vget_high_s32(__s0_238), __p1_238)); \
-  __ret_238; \
-})
-#else
-#define vshll_high_n_s32(__p0_239, __p1_239) __extension__ ({ \
-  int32x4_t __s0_239 = __p0_239; \
-  int32x4_t __rev0_239;  __rev0_239 = __builtin_shufflevector(__s0_239, __s0_239, 3, 2, 1, 0); \
-  int64x2_t __ret_239; \
-  __ret_239 = (int64x2_t)(__noswap_vshll_n_s32(__noswap_vget_high_s32(__rev0_239), __p1_239)); \
-  __ret_239 = __builtin_shufflevector(__ret_239, __ret_239, 1, 0); \
-  __ret_239; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshll_high_n_s16(__p0_240, __p1_240) __extension__ ({ \
-  int16x8_t __s0_240 = __p0_240; \
-  int32x4_t __ret_240; \
-  __ret_240 = (int32x4_t)(vshll_n_s16(vget_high_s16(__s0_240), __p1_240)); \
-  __ret_240; \
-})
-#else
-#define vshll_high_n_s16(__p0_241, __p1_241) __extension__ ({ \
-  int16x8_t __s0_241 = __p0_241; \
-  int16x8_t __rev0_241;  __rev0_241 = __builtin_shufflevector(__s0_241, __s0_241, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int32x4_t __ret_241; \
-  __ret_241 = (int32x4_t)(__noswap_vshll_n_s16(__noswap_vget_high_s16(__rev0_241), __p1_241)); \
-  __ret_241 = __builtin_shufflevector(__ret_241, __ret_241, 3, 2, 1, 0); \
-  __ret_241; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_u64(__p0, __p1) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vshrd_n_u64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#else
-#define vshrd_n_s64(__p0, __p1) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vshrd_n_s64(__s0, __p1); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u32(__p0_242, __p1_242, __p2_242) __extension__ ({ \
-  uint16x4_t __s0_242 = __p0_242; \
-  uint32x4_t __s1_242 = __p1_242; \
-  uint16x8_t __ret_242; \
-  __ret_242 = (uint16x8_t)(vcombine_u16((uint16x4_t)(__s0_242), (uint16x4_t)(vshrn_n_u32(__s1_242, __p2_242)))); \
-  __ret_242; \
-})
-#else
-#define vshrn_high_n_u32(__p0_243, __p1_243, __p2_243) __extension__ ({ \
-  uint16x4_t __s0_243 = __p0_243; \
-  uint32x4_t __s1_243 = __p1_243; \
-  uint16x4_t __rev0_243;  __rev0_243 = __builtin_shufflevector(__s0_243, __s0_243, 3, 2, 1, 0); \
-  uint32x4_t __rev1_243;  __rev1_243 = __builtin_shufflevector(__s1_243, __s1_243, 3, 2, 1, 0); \
-  uint16x8_t __ret_243; \
-  __ret_243 = (uint16x8_t)(__noswap_vcombine_u16((uint16x4_t)(__rev0_243), (uint16x4_t)(__noswap_vshrn_n_u32(__rev1_243, __p2_243)))); \
-  __ret_243 = __builtin_shufflevector(__ret_243, __ret_243, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_243; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u64(__p0_244, __p1_244, __p2_244) __extension__ ({ \
-  uint32x2_t __s0_244 = __p0_244; \
-  uint64x2_t __s1_244 = __p1_244; \
-  uint32x4_t __ret_244; \
-  __ret_244 = (uint32x4_t)(vcombine_u32((uint32x2_t)(__s0_244), (uint32x2_t)(vshrn_n_u64(__s1_244, __p2_244)))); \
-  __ret_244; \
-})
-#else
-#define vshrn_high_n_u64(__p0_245, __p1_245, __p2_245) __extension__ ({ \
-  uint32x2_t __s0_245 = __p0_245; \
-  uint64x2_t __s1_245 = __p1_245; \
-  uint32x2_t __rev0_245;  __rev0_245 = __builtin_shufflevector(__s0_245, __s0_245, 1, 0); \
-  uint64x2_t __rev1_245;  __rev1_245 = __builtin_shufflevector(__s1_245, __s1_245, 1, 0); \
-  uint32x4_t __ret_245; \
-  __ret_245 = (uint32x4_t)(__noswap_vcombine_u32((uint32x2_t)(__rev0_245), (uint32x2_t)(__noswap_vshrn_n_u64(__rev1_245, __p2_245)))); \
-  __ret_245 = __builtin_shufflevector(__ret_245, __ret_245, 3, 2, 1, 0); \
-  __ret_245; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_u16(__p0_246, __p1_246, __p2_246) __extension__ ({ \
-  uint8x8_t __s0_246 = __p0_246; \
-  uint16x8_t __s1_246 = __p1_246; \
-  uint8x16_t __ret_246; \
-  __ret_246 = (uint8x16_t)(vcombine_u8((uint8x8_t)(__s0_246), (uint8x8_t)(vshrn_n_u16(__s1_246, __p2_246)))); \
-  __ret_246; \
-})
-#else
-#define vshrn_high_n_u16(__p0_247, __p1_247, __p2_247) __extension__ ({ \
-  uint8x8_t __s0_247 = __p0_247; \
-  uint16x8_t __s1_247 = __p1_247; \
-  uint8x8_t __rev0_247;  __rev0_247 = __builtin_shufflevector(__s0_247, __s0_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint16x8_t __rev1_247;  __rev1_247 = __builtin_shufflevector(__s1_247, __s1_247, 7, 6, 5, 4, 3, 2, 1, 0); \
-  uint8x16_t __ret_247; \
-  __ret_247 = (uint8x16_t)(__noswap_vcombine_u8((uint8x8_t)(__rev0_247), (uint8x8_t)(__noswap_vshrn_n_u16(__rev1_247, __p2_247)))); \
-  __ret_247 = __builtin_shufflevector(__ret_247, __ret_247, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_247; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s32(__p0_248, __p1_248, __p2_248) __extension__ ({ \
-  int16x4_t __s0_248 = __p0_248; \
-  int32x4_t __s1_248 = __p1_248; \
-  int16x8_t __ret_248; \
-  __ret_248 = (int16x8_t)(vcombine_s16((int16x4_t)(__s0_248), (int16x4_t)(vshrn_n_s32(__s1_248, __p2_248)))); \
-  __ret_248; \
-})
-#else
-#define vshrn_high_n_s32(__p0_249, __p1_249, __p2_249) __extension__ ({ \
-  int16x4_t __s0_249 = __p0_249; \
-  int32x4_t __s1_249 = __p1_249; \
-  int16x4_t __rev0_249;  __rev0_249 = __builtin_shufflevector(__s0_249, __s0_249, 3, 2, 1, 0); \
-  int32x4_t __rev1_249;  __rev1_249 = __builtin_shufflevector(__s1_249, __s1_249, 3, 2, 1, 0); \
-  int16x8_t __ret_249; \
-  __ret_249 = (int16x8_t)(__noswap_vcombine_s16((int16x4_t)(__rev0_249), (int16x4_t)(__noswap_vshrn_n_s32(__rev1_249, __p2_249)))); \
-  __ret_249 = __builtin_shufflevector(__ret_249, __ret_249, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_249; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s64(__p0_250, __p1_250, __p2_250) __extension__ ({ \
-  int32x2_t __s0_250 = __p0_250; \
-  int64x2_t __s1_250 = __p1_250; \
-  int32x4_t __ret_250; \
-  __ret_250 = (int32x4_t)(vcombine_s32((int32x2_t)(__s0_250), (int32x2_t)(vshrn_n_s64(__s1_250, __p2_250)))); \
-  __ret_250; \
-})
-#else
-#define vshrn_high_n_s64(__p0_251, __p1_251, __p2_251) __extension__ ({ \
-  int32x2_t __s0_251 = __p0_251; \
-  int64x2_t __s1_251 = __p1_251; \
-  int32x2_t __rev0_251;  __rev0_251 = __builtin_shufflevector(__s0_251, __s0_251, 1, 0); \
-  int64x2_t __rev1_251;  __rev1_251 = __builtin_shufflevector(__s1_251, __s1_251, 1, 0); \
-  int32x4_t __ret_251; \
-  __ret_251 = (int32x4_t)(__noswap_vcombine_s32((int32x2_t)(__rev0_251), (int32x2_t)(__noswap_vshrn_n_s64(__rev1_251, __p2_251)))); \
-  __ret_251 = __builtin_shufflevector(__ret_251, __ret_251, 3, 2, 1, 0); \
-  __ret_251; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vshrn_high_n_s16(__p0_252, __p1_252, __p2_252) __extension__ ({ \
-  int8x8_t __s0_252 = __p0_252; \
-  int16x8_t __s1_252 = __p1_252; \
-  int8x16_t __ret_252; \
-  __ret_252 = (int8x16_t)(vcombine_s8((int8x8_t)(__s0_252), (int8x8_t)(vshrn_n_s16(__s1_252, __p2_252)))); \
-  __ret_252; \
-})
-#else
-#define vshrn_high_n_s16(__p0_253, __p1_253, __p2_253) __extension__ ({ \
-  int8x8_t __s0_253 = __p0_253; \
-  int16x8_t __s1_253 = __p1_253; \
-  int8x8_t __rev0_253;  __rev0_253 = __builtin_shufflevector(__s0_253, __s0_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16x8_t __rev1_253;  __rev1_253 = __builtin_shufflevector(__s1_253, __s1_253, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int8x16_t __ret_253; \
-  __ret_253 = (int8x16_t)(__noswap_vcombine_s8((int8x8_t)(__rev0_253), (int8x8_t)(__noswap_vshrn_n_s16(__rev1_253, __p2_253)))); \
-  __ret_253 = __builtin_shufflevector(__ret_253, __ret_253, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_253; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vslid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vslid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vslid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vslid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#else
-#define vsli_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsli_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
-  __ret; \
-})
-#else
-#define vsliq_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsliq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint8_t vsqaddb_u8(uint8_t __p0, int8_t __p1) {
-  uint8_t __ret;
-  __ret = (uint8_t) __builtin_neon_vsqaddb_u8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint32_t vsqadds_u32(uint32_t __p0, int32_t __p1) {
-  uint32_t __ret;
-  __ret = (uint32_t) __builtin_neon_vsqadds_u32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vsqaddd_u64(uint64_t __p0, int64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsqaddd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint16_t vsqaddh_u16(uint16_t __p0, int16_t __p1) {
-  uint16_t __ret;
-  __ret = (uint16_t) __builtin_neon_vsqaddh_u16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 48);
-  return __ret;
-}
-#else
-__ai uint8x16_t vsqaddq_u8(uint8x16_t __p0, int8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = (uint8x16_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 48);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 50);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsqaddq_u32(uint32x4_t __p0, int32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 50);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vsqaddq_u64(uint64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 49);
-  return __ret;
-}
-#else
-__ai uint16x8_t vsqaddq_u16(uint16x8_t __p0, int16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t) __builtin_neon_vsqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 49);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vsqadd_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 16);
-  return __ret;
-}
-#else
-__ai uint8x8_t vsqadd_u8(uint8x8_t __p0, int8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = (uint8x8_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 16);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vsqadd_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 18);
-  return __ret;
-}
-#else
-__ai uint32x2_t vsqadd_u32(uint32x2_t __p0, int32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = (uint32x2_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 18);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vsqadd_u64(uint64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 17);
-  return __ret;
-}
-#else
-__ai uint16x4_t vsqadd_u16(uint16x4_t __p0, int16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = (uint16x4_t) __builtin_neon_vsqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 17);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 42);
-  return __ret;
-}
-#else
-__ai float64x2_t vsqrtq_f64(float64x2_t __p0) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __ret;
-  __ret = (float64x2_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 42);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__p0, 41);
-  return __ret;
-}
-#else
-__ai float32x4_t vsqrtq_f32(float32x4_t __p0) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = (float32x4_t) __builtin_neon_vsqrtq_v((int8x16_t)__rev0, 41);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#else
-__ai float64x1_t vsqrt_f64(float64x1_t __p0) {
-  float64x1_t __ret;
-  __ret = (float64x1_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 10);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vsqrt_f32(float32x2_t __p0) {
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__p0, 9);
-  return __ret;
-}
-#else
-__ai float32x2_t vsqrt_f32(float32x2_t __p0) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __ret;
-  __ret = (float32x2_t) __builtin_neon_vsqrt_v((int8x8_t)__rev0, 9);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vsrad_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrad_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vsrad_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrad_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vsrid_n_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64_t __s0 = __p0; \
-  uint64_t __s1 = __p1; \
-  uint64_t __ret; \
-  __ret = (uint64_t) __builtin_neon_vsrid_n_u64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#else
-#define vsrid_n_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64_t __s0 = __p0; \
-  int64_t __s1 = __p1; \
-  int64_t __ret; \
-  __ret = (int64_t) __builtin_neon_vsrid_n_s64(__s0, __s1, __p2); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#else
-#define vsri_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s0 = __p0; \
-  poly64x1_t __s1 = __p1; \
-  poly64x1_t __ret; \
-  __ret = (poly64x1_t) __builtin_neon_vsri_n_v((int8x8_t)__s0, (int8x8_t)__s1, __p2, 6); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__s0, (int8x16_t)__s1, __p2, 38); \
-  __ret; \
-})
-#else
-#define vsriq_n_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s0 = __p0; \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  poly64x2_t __ret; \
-  __ret = (poly64x2_t) __builtin_neon_vsriq_n_v((int8x16_t)__rev0, (int8x16_t)__rev1, __p2, 38); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
-})
-#else
-#define vst1_p64(__p0, __p1) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 38); \
-})
-#else
-#define vst1q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__s1, 42); \
-})
-#else
-#define vst1q_f64(__p0, __p1) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_v(__p0, (int8x16_t)__rev1, 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
-})
-#else
-#define vst1_f64(__p0, __p1) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_v(__p0, (int8x8_t)__s1, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-})
-#else
-#define vst1_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 38); \
-})
-#else
-#define vst1q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2_t __s1 = __p1; \
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__s1, __p2, 42); \
-})
-#else
-#define vst1q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2_t __s1 = __p1; \
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  __builtin_neon_vst1q_lane_v(__p0, (int8x16_t)__rev1, __p2, 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-})
-#else
-#define vst1_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1_t __s1 = __p1; \
-  __builtin_neon_vst1_lane_v(__p0, (int8x8_t)__s1, __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#else
-#define vst1_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
-})
-#else
-#define vst1q_p64_x2(__p0, __p1) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  poly64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst1q_x2_v(__p0, __s1.val[0], __s1.val[1], 42); \
-})
-#else
-#define vst1q_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  float64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst1q_x2_v(__p0, __rev1.val[0], __rev1.val[1], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#else
-#define vst1_f64_x2(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst1_x2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#else
-#define vst1_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
-})
-#else
-#define vst1q_p64_x3(__p0, __p1) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  poly64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst1q_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 42); \
-})
-#else
-#define vst1q_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  float64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst1q_x3_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#else
-#define vst1_f64_x3(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst1_x3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#else
-#define vst1_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
-})
-#else
-#define vst1q_p64_x4(__p0, __p1) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  poly64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst1q_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 42); \
-})
-#else
-#define vst1q_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  float64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst1q_x4_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#else
-#define vst1_f64_x4(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst1_x4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#else
-#define vst2_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 38); \
-})
-#else
-#define vst2q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  poly64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], 51); \
-})
-#else
-#define vst2q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  uint64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 42); \
-})
-#else
-#define vst2q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  float64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_v(__p0, __s1.val[0], __s1.val[1], 35); \
-})
-#else
-#define vst2q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  int64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_v(__p0, __rev1.val[0], __rev1.val[1], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_f64(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#else
-#define vst2_f64(__p0, __p1) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_v(__p0, __s1.val[0], __s1.val[1], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-})
-#else
-#define vst2_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 36); \
-})
-#else
-#define vst2q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x2_t __s1 = __p1; \
-  poly8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 38); \
-})
-#else
-#define vst2q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x2_t __s1 = __p1; \
-  poly64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 48); \
-})
-#else
-#define vst2q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x2_t __s1 = __p1; \
-  uint8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 51); \
-})
-#else
-#define vst2q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x2_t __s1 = __p1; \
-  uint64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], __p2, 32); \
-})
-#else
-#define vst2q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x2_t __s1 = __p1; \
-  int8x16x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], __p2, 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 42); \
-})
-#else
-#define vst2q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x2_t __s1 = __p1; \
-  float64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  __builtin_neon_vst2q_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 35); \
-})
-#else
-#define vst2q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x2_t __s1 = __p1; \
-  int64x2x2_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __builtin_neon_vst2q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __p2, 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-})
-#else
-#define vst2_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
-})
-#else
-#define vst2_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
-})
-#else
-#define vst2_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x2_t __s1 = __p1; \
-  __builtin_neon_vst2_lane_v(__p0, __s1.val[0], __s1.val[1], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#else
-#define vst3_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 38); \
-})
-#else
-#define vst3q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  poly64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], 51); \
-})
-#else
-#define vst3q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  uint64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 42); \
-})
-#else
-#define vst3q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  float64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 35); \
-})
-#else
-#define vst3q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  int64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_f64(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#else
-#define vst3_f64(__p0, __p1) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-})
-#else
-#define vst3_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 36); \
-})
-#else
-#define vst3q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x3_t __s1 = __p1; \
-  poly8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 38); \
-})
-#else
-#define vst3q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x3_t __s1 = __p1; \
-  poly64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 48); \
-})
-#else
-#define vst3q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x3_t __s1 = __p1; \
-  uint8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 51); \
-})
-#else
-#define vst3q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x3_t __s1 = __p1; \
-  uint64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], __p2, 32); \
-})
-#else
-#define vst3q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x3_t __s1 = __p1; \
-  int8x16x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], __p2, 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 42); \
-})
-#else
-#define vst3q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x3_t __s1 = __p1; \
-  float64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  __builtin_neon_vst3q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 35); \
-})
-#else
-#define vst3q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x3_t __s1 = __p1; \
-  int64x2x3_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __builtin_neon_vst3q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __p2, 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-})
-#else
-#define vst3_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-})
-#else
-#define vst3_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-})
-#else
-#define vst3_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x3_t __s1 = __p1; \
-  __builtin_neon_vst3_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#else
-#define vst4_p64(__p0, __p1) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 38); \
-})
-#else
-#define vst4q_p64(__p0, __p1) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  poly64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], 51); \
-})
-#else
-#define vst4q_u64(__p0, __p1) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  uint64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 42); \
-})
-#else
-#define vst4q_f64(__p0, __p1) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  float64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 35); \
-})
-#else
-#define vst4q_s64(__p0, __p1) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  int64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_f64(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#else
-#define vst4_f64(__p0, __p1) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-})
-#else
-#define vst4_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 6); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 36); \
-})
-#else
-#define vst4q_lane_p8(__p0, __p1, __p2) __extension__ ({ \
-  poly8x16x4_t __s1 = __p1; \
-  poly8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 36); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 38); \
-})
-#else
-#define vst4q_lane_p64(__p0, __p1, __p2) __extension__ ({ \
-  poly64x2x4_t __s1 = __p1; \
-  poly64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 38); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 48); \
-})
-#else
-#define vst4q_lane_u8(__p0, __p1, __p2) __extension__ ({ \
-  uint8x16x4_t __s1 = __p1; \
-  uint8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 48); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 51); \
-})
-#else
-#define vst4q_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x2x4_t __s1 = __p1; \
-  uint64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 51); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__s1.val[0], (int8x16_t)__s1.val[1], (int8x16_t)__s1.val[2], (int8x16_t)__s1.val[3], __p2, 32); \
-})
-#else
-#define vst4q_lane_s8(__p0, __p1, __p2) __extension__ ({ \
-  int8x16x4_t __s1 = __p1; \
-  int8x16x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, (int8x16_t)__rev1.val[0], (int8x16_t)__rev1.val[1], (int8x16_t)__rev1.val[2], (int8x16_t)__rev1.val[3], __p2, 32); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 42); \
-})
-#else
-#define vst4q_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x2x4_t __s1 = __p1; \
-  float64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 42); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  __builtin_neon_vst4q_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 35); \
-})
-#else
-#define vst4q_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x2x4_t __s1 = __p1; \
-  int64x2x4_t __rev1; \
-  __rev1.val[0] = __builtin_shufflevector(__s1.val[0], __s1.val[0], 1, 0); \
-  __rev1.val[1] = __builtin_shufflevector(__s1.val[1], __s1.val[1], 1, 0); \
-  __rev1.val[2] = __builtin_shufflevector(__s1.val[2], __s1.val[2], 1, 0); \
-  __rev1.val[3] = __builtin_shufflevector(__s1.val[3], __s1.val[3], 1, 0); \
-  __builtin_neon_vst4q_lane_v(__p0, __rev1.val[0], __rev1.val[1], __rev1.val[2], __rev1.val[3], __p2, 35); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-})
-#else
-#define vst4_lane_u64(__p0, __p1, __p2) __extension__ ({ \
-  uint64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, (int8x8_t)__s1.val[0], (int8x8_t)__s1.val[1], (int8x8_t)__s1.val[2], (int8x8_t)__s1.val[3], __p2, 19); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-})
-#else
-#define vst4_lane_f64(__p0, __p1, __p2) __extension__ ({ \
-  float64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 10); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-})
-#else
-#define vst4_lane_s64(__p0, __p1, __p2) __extension__ ({ \
-  int64x1x4_t __s1 = __p1; \
-  __builtin_neon_vst4_lane_v(__p0, __s1.val[0], __s1.val[1], __s1.val[2], __s1.val[3], __p2, 3); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vstrq_p128(__p0, __p1) __extension__ ({ \
-  poly128_t __s1 = __p1; \
-  __builtin_neon_vstrq_p128(__p0, __s1); \
-})
-#else
-#define vstrq_p128(__p0, __p1) __extension__ ({ \
-  poly128_t __s1 = __p1; \
-  __builtin_neon_vstrq_p128(__p0, __s1); \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vsubd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vsubd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vsubd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vsubd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float64x2_t vsubq_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __rev0 - __rev1;
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#else
-__ai float64x1_t vsub_f64(float64x1_t __p0, float64x1_t __p1) {
-  float64x1_t __ret;
-  __ret = __p0 - __p1;
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x8_t __ret;
-  __ret = vcombine_u16(__p0, vsubhn_u32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubhn_high_u32(uint16x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vcombine_u16(__rev0, __noswap_vsubhn_u32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x4_t __ret;
-  __ret = vcombine_u32(__p0, vsubhn_u64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubhn_high_u64(uint32x2_t __p0, uint64x2_t __p1, uint64x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vcombine_u32(__rev0, __noswap_vsubhn_u64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x16_t __ret;
-  __ret = vcombine_u8(__p0, vsubhn_u16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai uint8x16_t vsubhn_high_u16(uint8x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __noswap_vcombine_u8(__rev0, __noswap_vsubhn_u16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x8_t __ret;
-  __ret = vcombine_s16(__p0, vsubhn_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vsubhn_high_s32(int16x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vcombine_s16(__rev0, __noswap_vsubhn_s32(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x4_t __ret;
-  __ret = vcombine_s32(__p0, vsubhn_s64(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vsubhn_high_s64(int32x2_t __p0, int64x2_t __p1, int64x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vcombine_s32(__rev0, __noswap_vsubhn_s64(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x16_t __ret;
-  __ret = vcombine_s8(__p0, vsubhn_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int8x16_t vsubhn_high_s16(int8x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __noswap_vcombine_s8(__rev0, __noswap_vsubhn_s16(__rev1, __rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = vmovl_high_u8(__p0) - vmovl_high_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmovl_high_u8(__rev0) - __noswap_vmovl_high_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmovl_high_u32(__p0) - vmovl_high_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vsubl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmovl_high_u32(__rev0) - __noswap_vmovl_high_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmovl_high_u16(__p0) - vmovl_high_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmovl_high_u16(__rev0) - __noswap_vmovl_high_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = vmovl_high_s8(__p0) - vmovl_high_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vsubl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmovl_high_s8(__rev0) - __noswap_vmovl_high_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = vmovl_high_s32(__p0) - vmovl_high_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vsubl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmovl_high_s32(__rev0) - __noswap_vmovl_high_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = vmovl_high_s16(__p0) - vmovl_high_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vsubl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmovl_high_s16(__rev0) - __noswap_vmovl_high_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 - vmovl_high_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vsubw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 - vmovl_high_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vsubw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 - vmovl_high_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vsubw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 - vmovl_high_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vsubw_high_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 - vmovl_high_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vsubw_high_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 - vmovl_high_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vsubw_high_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __noswap_vmovl_high_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtrn1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai poly16x4_t vtrn1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  return __ret;
-}
-#else
-__ai poly8x16_t vtrn1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai poly64x2_t vtrn1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai poly16x8_t vtrn1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  return __ret;
-}
-#else
-__ai uint8x16_t vtrn1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai uint32x4_t vtrn1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vtrn1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai uint16x8_t vtrn1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  return __ret;
-}
-#else
-__ai int8x16_t vtrn1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 2, 18, 4, 20, 6, 22, 8, 24, 10, 26, 12, 28, 14, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float64x2_t vtrn1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai float32x4_t vtrn1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai int32x4_t vtrn1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int64x2_t vtrn1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai int16x8_t vtrn1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtrn1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint32x2_t vtrn1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai uint16x4_t vtrn1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 2, 10, 4, 12, 6, 14);
-  return __ret;
-}
-#else
-__ai int8x8_t vtrn1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 2, 10, 4, 12, 6, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float32x2_t vtrn1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vtrn1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 2, 6);
-  return __ret;
-}
-#else
-__ai int16x4_t vtrn1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 2, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai poly8x8_t vtrn2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai poly16x4_t vtrn2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  return __ret;
-}
-#else
-__ai poly8x16_t vtrn2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai poly64x2_t vtrn2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai poly16x8_t vtrn2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  return __ret;
-}
-#else
-__ai uint8x16_t vtrn2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai uint32x4_t vtrn2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint64x2_t vtrn2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai uint16x8_t vtrn2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  return __ret;
-}
-#else
-__ai int8x16_t vtrn2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 17, 3, 19, 5, 21, 7, 23, 9, 25, 11, 27, 13, 29, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float64x2_t vtrn2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai float32x4_t vtrn2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai int32x4_t vtrn2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int64x2_t vtrn2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai int16x8_t vtrn2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai uint8x8_t vtrn2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint32x2_t vtrn2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai uint16x4_t vtrn2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 9, 3, 11, 5, 13, 7, 15);
-  return __ret;
-}
-#else
-__ai int8x8_t vtrn2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 9, 3, 11, 5, 13, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float32x2_t vtrn2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int32x2_t vtrn2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 5, 3, 7);
-  return __ret;
-}
-#else
-__ai int16x4_t vtrn2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 5, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vtst_p64(poly64x1_t __p0, poly64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vtstq_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vtstq_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__p0, (int8x16_t)__p1, 51);
-  return __ret;
-}
-#else
-__ai uint64x2_t vtstq_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t) __builtin_neon_vtstq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 51);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vtst_u64(uint64x1_t __p0, uint64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#else
-__ai uint64x1_t vtst_s64(int64x1_t __p0, int64x1_t __p1) {
-  uint64x1_t __ret;
-  __ret = (uint64x1_t) __builtin_neon_vtst_v((int8x8_t)__p0, (int8x8_t)__p1, 19);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai uint64_t vtstd_u64(uint64_t __p0, uint64_t __p1) {
-  uint64_t __ret;
-  __ret = (uint64_t) __builtin_neon_vtstd_u64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vtstd_s64(int64_t __p0, int64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vtstd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int8_t vuqaddb_s8(int8_t __p0, uint8_t __p1) {
-  int8_t __ret;
-  __ret = (int8_t) __builtin_neon_vuqaddb_s8(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int32_t vuqadds_s32(int32_t __p0, uint32_t __p1) {
-  int32_t __ret;
-  __ret = (int32_t) __builtin_neon_vuqadds_s32(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int64_t vuqaddd_s64(int64_t __p0, uint64_t __p1) {
-  int64_t __ret;
-  __ret = (int64_t) __builtin_neon_vuqaddd_s64(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#else
-__ai int16_t vuqaddh_s16(int16_t __p0, uint16_t __p1) {
-  int16_t __ret;
-  __ret = (int16_t) __builtin_neon_vuqaddh_s16(__p0, __p1);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 32);
-  return __ret;
-}
-#else
-__ai int8x16_t vuqaddq_s8(int8x16_t __p0, uint8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = (int8x16_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 32);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vuqaddq_s32(int32x4_t __p0, uint32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 34);
-  return __ret;
-}
-#else
-__ai int32x4_t vuqaddq_s32(int32x4_t __p0, uint32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 34);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vuqaddq_s64(int64x2_t __p0, uint64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 35);
-  return __ret;
-}
-#else
-__ai int64x2_t vuqaddq_s64(int64x2_t __p0, uint64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 35);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vuqaddq_s16(int16x8_t __p0, uint16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__p0, (int8x16_t)__p1, 33);
-  return __ret;
-}
-#else
-__ai int16x8_t vuqaddq_s16(int16x8_t __p0, uint16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t) __builtin_neon_vuqaddq_v((int8x16_t)__rev0, (int8x16_t)__rev1, 33);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vuqadd_s8(int8x8_t __p0, uint8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 0);
-  return __ret;
-}
-#else
-__ai int8x8_t vuqadd_s8(int8x8_t __p0, uint8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = (int8x8_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 0);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vuqadd_s32(int32x2_t __p0, uint32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vuqadd_s32(int32x2_t __p0, uint32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = (int32x2_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#else
-__ai int64x1_t vuqadd_s64(int64x1_t __p0, uint64x1_t __p1) {
-  int64x1_t __ret;
-  __ret = (int64x1_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 3);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__p0, (int8x8_t)__p1, 1);
-  return __ret;
-}
-#else
-__ai int16x4_t vuqadd_s16(int16x4_t __p0, uint16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = (int16x4_t) __builtin_neon_vuqadd_v((int8x8_t)__rev0, (int8x8_t)__rev1, 1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai poly8x8_t vuzp1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai poly16x4_t vuzp1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  return __ret;
-}
-#else
-__ai poly8x16_t vuzp1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai poly64x2_t vuzp1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai poly16x8_t vuzp1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  return __ret;
-}
-#else
-__ai uint8x16_t vuzp1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai uint32x4_t vuzp1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vuzp1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai uint16x8_t vuzp1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  return __ret;
-}
-#else
-__ai int8x16_t vuzp1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float64x2_t vuzp1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai float32x4_t vuzp1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai int32x4_t vuzp1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int64x2_t vuzp1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai int16x8_t vuzp1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai uint8x8_t vuzp1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint32x2_t vuzp1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai uint16x4_t vuzp1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6, 8, 10, 12, 14);
-  return __ret;
-}
-#else
-__ai int8x8_t vuzp1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6, 8, 10, 12, 14);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float32x2_t vuzp1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vuzp1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2, 4, 6);
-  return __ret;
-}
-#else
-__ai int16x4_t vuzp1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2, 4, 6);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai poly8x8_t vuzp2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai poly16x4_t vuzp2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  return __ret;
-}
-#else
-__ai poly8x16_t vuzp2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai poly64x2_t vuzp2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai poly16x8_t vuzp2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  return __ret;
-}
-#else
-__ai uint8x16_t vuzp2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai uint32x4_t vuzp2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint64x2_t vuzp2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai uint16x8_t vuzp2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  return __ret;
-}
-#else
-__ai int8x16_t vuzp2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float64x2_t vuzp2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai float32x4_t vuzp2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai int32x4_t vuzp2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int64x2_t vuzp2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai int16x8_t vuzp2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai uint8x8_t vuzp2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint32x2_t vuzp2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai uint16x4_t vuzp2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7, 9, 11, 13, 15);
-  return __ret;
-}
-#else
-__ai int8x8_t vuzp2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7, 9, 11, 13, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float32x2_t vuzp2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int32x2_t vuzp2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3, 5, 7);
-  return __ret;
-}
-#else
-__ai int16x4_t vuzp2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3, 5, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai poly8x8_t vzip1_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai poly16x4_t vzip1_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  return __ret;
-}
-#else
-__ai poly8x16_t vzip1q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai poly64x2_t vzip1q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai poly16x8_t vzip1q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  return __ret;
-}
-#else
-__ai uint8x16_t vzip1q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai uint32x4_t vzip1q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vzip1q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai uint16x8_t vzip1q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  return __ret;
-}
-#else
-__ai int8x16_t vzip1q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 16, 1, 17, 2, 18, 3, 19, 4, 20, 5, 21, 6, 22, 7, 23);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float64x2_t vzip1q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai float32x4_t vzip1q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai int32x4_t vzip1q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int64x2_t vzip1q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai int16x8_t vzip1q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai uint8x8_t vzip1_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai uint32x2_t vzip1_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai uint16x4_t vzip1_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 8, 1, 9, 2, 10, 3, 11);
-  return __ret;
-}
-#else
-__ai int8x8_t vzip1_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 8, 1, 9, 2, 10, 3, 11);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai float32x2_t vzip1_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 2);
-  return __ret;
-}
-#else
-__ai int32x2_t vzip1_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 0, 4, 1, 5);
-  return __ret;
-}
-#else
-__ai int16x4_t vzip1_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 0, 4, 1, 5);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai poly8x8_t vzip2_p8(poly8x8_t __p0, poly8x8_t __p1) {
-  poly8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai poly16x4_t vzip2_p16(poly16x4_t __p0, poly16x4_t __p1) {
-  poly16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  poly16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  poly16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  return __ret;
-}
-#else
-__ai poly8x16_t vzip2q_p8(poly8x16_t __p0, poly8x16_t __p1) {
-  poly8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai poly64x2_t vzip2q_p64(poly64x2_t __p0, poly64x2_t __p1) {
-  poly64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  poly64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  poly64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai poly16x8_t vzip2q_p16(poly16x8_t __p0, poly16x8_t __p1) {
-  poly16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  poly16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  return __ret;
-}
-#else
-__ai uint8x16_t vzip2q_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai uint32x4_t vzip2q_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint64x2_t vzip2q_u64(uint64x2_t __p0, uint64x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai uint16x8_t vzip2q_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  return __ret;
-}
-#else
-__ai int8x16_t vzip2q_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 8, 24, 9, 25, 10, 26, 11, 27, 12, 28, 13, 29, 14, 30, 15, 31);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float64x2_t vzip2q_f64(float64x2_t __p0, float64x2_t __p1) {
-  float64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai float32x4_t vzip2q_f32(float32x4_t __p0, float32x4_t __p1) {
-  float32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  float32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  float32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai int32x4_t vzip2q_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int64x2_t vzip2q_s64(int64x2_t __p0, int64x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int64x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai int16x8_t vzip2q_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai uint8x8_t vzip2_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai uint32x2_t vzip2_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai uint16x4_t vzip2_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 4, 12, 5, 13, 6, 14, 7, 15);
-  return __ret;
-}
-#else
-__ai int8x8_t vzip2_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 4, 12, 5, 13, 6, 14, 7, 15);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai float32x2_t vzip2_f32(float32x2_t __p0, float32x2_t __p1) {
-  float32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  float32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  float32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 1, 3);
-  return __ret;
-}
-#else
-__ai int32x2_t vzip2_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 1, 3);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__p0, __p1, 2, 6, 3, 7);
-  return __ret;
-}
-#else
-__ai int16x4_t vzip2_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __builtin_shufflevector(__rev0, __rev1, 2, 6, 3, 7);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __ret;
-  __ret = __p0 + vabdq_u8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint8x16_t vabaq_u8(uint8x16_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __ret;
-  __ret = __rev0 + __noswap_vabdq_u8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + vabdq_u32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vabaq_u32(uint32x4_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vabdq_u32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + vabdq_u16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint16x8_t vabaq_u16(uint16x8_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __noswap_vabdq_u16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __ret;
-  __ret = __p0 + vabdq_s8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int8x16_t vabaq_s8(int8x16_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __ret;
-  __ret = __rev0 + __noswap_vabdq_s8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + vabdq_s32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vabaq_s32(int32x4_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vabdq_s32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + vabdq_s16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int16x8_t vabaq_s16(int16x8_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __noswap_vabdq_s16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __ret;
-  __ret = __p0 + vabd_u8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint8x8_t vaba_u8(uint8x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __ret;
-  __ret = __rev0 + __noswap_vabd_u8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __ret;
-  __ret = __p0 + vabd_u32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint32x2_t vaba_u32(uint32x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint32x2_t __ret;
-  __ret = __rev0 + __noswap_vabd_u32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __ret;
-  __ret = __p0 + vabd_u16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint16x4_t vaba_u16(uint16x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint16x4_t __ret;
-  __ret = __rev0 + __noswap_vabd_u16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __ret;
-  __ret = __p0 + vabd_s8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int8x8_t vaba_s8(int8x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __ret;
-  __ret = __rev0 + __noswap_vabd_s8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __ret;
-  __ret = __p0 + vabd_s32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int32x2_t vaba_s32(int32x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int32x2_t __ret;
-  __ret = __rev0 + __noswap_vabd_s32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __ret;
-  __ret = __p0 + vabd_s16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int16x4_t vaba_s16(int16x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int16x4_t __ret;
-  __ret = __rev0 + __noswap_vabd_s16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(vmovl_u8((uint8x8_t)(vabd_u8(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai uint16x8_t vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vabdl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = (uint16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_u8(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(vmovl_u32((uint32x2_t)(vabd_u32(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai uint64x2_t vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vabdl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = (uint64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_u32(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(vmovl_u16((uint16x4_t)(vabd_u16(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai uint32x4_t vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vabdl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = (uint32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_u16(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(vmovl_u8((uint8x8_t)(vabd_s8(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai int16x8_t vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vabdl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = (int16x8_t)(__noswap_vmovl_u8((uint8x8_t)(__noswap_vabd_s8(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(vmovl_u32((uint32x2_t)(vabd_s32(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai int64x2_t vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vabdl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = (int64x2_t)(__noswap_vmovl_u32((uint32x2_t)(__noswap_vabd_s32(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(vmovl_u16((uint16x4_t)(vabd_s16(__p0, __p1))));
-  return __ret;
-}
-#else
-__ai int32x4_t vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__rev0, __rev1))));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vabdl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = (int32x4_t)(__noswap_vmovl_u16((uint16x4_t)(__noswap_vabd_s16(__p0, __p1))));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = vmovl_u8(__p0) + vmovl_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddl_u8(uint8x8_t __p0, uint8x8_t __p1) {
-  uint8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmovl_u8(__rev0) + __noswap_vmovl_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmovl_u32(__p0) + vmovl_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vaddl_u32(uint32x2_t __p0, uint32x2_t __p1) {
-  uint32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmovl_u32(__rev0) + __noswap_vmovl_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmovl_u16(__p0) + vmovl_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddl_u16(uint16x4_t __p0, uint16x4_t __p1) {
-  uint16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmovl_u16(__rev0) + __noswap_vmovl_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = vmovl_s8(__p0) + vmovl_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vaddl_s8(int8x8_t __p0, int8x8_t __p1) {
-  int8x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmovl_s8(__rev0) + __noswap_vmovl_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = vmovl_s32(__p0) + vmovl_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vaddl_s32(int32x2_t __p0, int32x2_t __p1) {
-  int32x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmovl_s32(__rev0) + __noswap_vmovl_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = vmovl_s16(__p0) + vmovl_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vaddl_s16(int16x4_t __p0, int16x4_t __p1) {
-  int16x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmovl_s16(__rev0) + __noswap_vmovl_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 + vmovl_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddw_u8(uint16x8_t __p0, uint8x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __noswap_vmovl_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 + vmovl_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vaddw_u32(uint64x2_t __p0, uint32x2_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __noswap_vmovl_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 + vmovl_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddw_u16(uint32x4_t __p0, uint16x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vmovl_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 + vmovl_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vaddw_s8(int16x8_t __p0, int8x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __noswap_vmovl_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 + vmovl_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vaddw_s32(int64x2_t __p0, int32x2_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __noswap_vmovl_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 + vmovl_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vaddw_s16(int32x4_t __p0, int16x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vmovl_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vget_lane_f16(__p0_254, __p1_254) __extension__ ({ \
-  float16x4_t __s0_254 = __p0_254; \
-  float16_t __ret_254; \
-float16x4_t __reint_254 = __s0_254; \
-int16_t __reint1_254 = vget_lane_s16(*(int16x4_t *) &__reint_254, __p1_254); \
-  __ret_254 = *(float16_t *) &__reint1_254; \
-  __ret_254; \
-})
-#else
-#define vget_lane_f16(__p0_255, __p1_255) __extension__ ({ \
-  float16x4_t __s0_255 = __p0_255; \
-  float16x4_t __rev0_255;  __rev0_255 = __builtin_shufflevector(__s0_255, __s0_255, 3, 2, 1, 0); \
-  float16_t __ret_255; \
-float16x4_t __reint_255 = __rev0_255; \
-int16_t __reint1_255 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_255, __p1_255); \
-  __ret_255 = *(float16_t *) &__reint1_255; \
-  __ret_255; \
-})
-#define __noswap_vget_lane_f16(__p0_256, __p1_256) __extension__ ({ \
-  float16x4_t __s0_256 = __p0_256; \
-  float16_t __ret_256; \
-float16x4_t __reint_256 = __s0_256; \
-int16_t __reint1_256 = __noswap_vget_lane_s16(*(int16x4_t *) &__reint_256, __p1_256); \
-  __ret_256 = *(float16_t *) &__reint1_256; \
-  __ret_256; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vgetq_lane_f16(__p0_257, __p1_257) __extension__ ({ \
-  float16x8_t __s0_257 = __p0_257; \
-  float16_t __ret_257; \
-float16x8_t __reint_257 = __s0_257; \
-int16_t __reint1_257 = vgetq_lane_s16(*(int16x8_t *) &__reint_257, __p1_257); \
-  __ret_257 = *(float16_t *) &__reint1_257; \
-  __ret_257; \
-})
-#else
-#define vgetq_lane_f16(__p0_258, __p1_258) __extension__ ({ \
-  float16x8_t __s0_258 = __p0_258; \
-  float16x8_t __rev0_258;  __rev0_258 = __builtin_shufflevector(__s0_258, __s0_258, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_258; \
-float16x8_t __reint_258 = __rev0_258; \
-int16_t __reint1_258 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_258, __p1_258); \
-  __ret_258 = *(float16_t *) &__reint1_258; \
-  __ret_258; \
-})
-#define __noswap_vgetq_lane_f16(__p0_259, __p1_259) __extension__ ({ \
-  float16x8_t __s0_259 = __p0_259; \
-  float16_t __ret_259; \
-float16x8_t __reint_259 = __s0_259; \
-int16_t __reint1_259 = __noswap_vgetq_lane_s16(*(int16x8_t *) &__reint_259, __p1_259); \
-  __ret_259 = *(float16_t *) &__reint1_259; \
-  __ret_259; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + vmull_u8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __noswap_vmull_u8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vmlal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + __noswap_vmull_u8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + vmull_u32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __noswap_vmull_u32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmlal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + __noswap_vmull_u32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + vmull_u16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vmull_u16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmlal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + __noswap_vmull_u16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + vmull_s8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int16x8_t vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __noswap_vmull_s8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vmlal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + __noswap_vmull_s8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + vmull_s32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __noswap_vmull_s32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmlal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + __noswap_vmull_s32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + vmull_s16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vmull_s16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmlal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + __noswap_vmull_s16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 + vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 + vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 + vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 + vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlal_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 + __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmlal_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmlal_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + vmull_s32(__p1, (int32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai int64x2_t vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmlal_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai int32x4_t vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmlal_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 - vmull_u8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 - __noswap_vmull_u8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vmlsl_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 - __noswap_vmull_u8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 - vmull_u32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 - __noswap_vmull_u32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmlsl_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 - __noswap_vmull_u32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - vmull_u16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __noswap_vmull_u16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmlsl_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - __noswap_vmull_u16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 - vmull_s8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int16x8_t vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 - __noswap_vmull_s8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vmlsl_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 - __noswap_vmull_s8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 - vmull_s32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 - __noswap_vmull_s32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmlsl_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 - __noswap_vmull_s32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - vmull_s16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __noswap_vmull_s16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmlsl_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - __noswap_vmull_s16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __ret; \
-  __ret = __s0 - vmull_u32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_lane_u32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint64x2_t __s0 = __p0; \
-  uint32x2_t __s1 = __p1; \
-  uint32x2_t __s2 = __p2; \
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  uint64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __ret; \
-  __ret = __s0 - vmull_u16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_lane_u16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  uint32x4_t __s0 = __p0; \
-  uint16x4_t __s1 = __p1; \
-  uint16x4_t __s2 = __p2; \
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  uint32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_u16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __ret; \
-  __ret = __s0 - vmull_s32(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_lane_s32(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int64x2_t __s0 = __p0; \
-  int32x2_t __s1 = __p1; \
-  int32x2_t __s2 = __p2; \
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 1, 0); \
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 1, 0); \
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 1, 0); \
-  int64x2_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s32(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __ret; \
-  __ret = __s0 - vmull_s16(__s1, __builtin_shufflevector(__s2, __s2, __p3, __p3, __p3, __p3)); \
-  __ret; \
-})
-#else
-#define vmlsl_lane_s16(__p0, __p1, __p2, __p3) __extension__ ({ \
-  int32x4_t __s0 = __p0; \
-  int16x4_t __s1 = __p1; \
-  int16x4_t __s2 = __p2; \
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__s0, __s0, 3, 2, 1, 0); \
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__s1, __s1, 3, 2, 1, 0); \
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__s2, __s2, 3, 2, 1, 0); \
-  int32x4_t __ret; \
-  __ret = __rev0 - __noswap_vmull_s16(__rev1, __builtin_shufflevector(__rev2, __rev2, __p3, __p3, __p3, __p3)); \
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0); \
-  __ret; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 - vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 - __noswap_vmull_u32(__rev1, (uint32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vmlsl_n_u32(uint64x2_t __p0, uint32x2_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 - __noswap_vmull_u32(__p1, (uint32x2_t) {__p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 - __noswap_vmull_u16(__rev1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vmlsl_n_u16(uint32x4_t __p0, uint16x4_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 - __noswap_vmull_u16(__p1, (uint16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 - vmull_s32(__p1, (int32x2_t) {__p2, __p2});
-  return __ret;
-}
-#else
-__ai int64x2_t vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 - __noswap_vmull_s32(__rev1, (int32x2_t) {__p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vmlsl_n_s32(int64x2_t __p0, int32x2_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 - __noswap_vmull_s32(__p1, (int32x2_t) {__p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 - __noswap_vmull_s16(__rev1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vmlsl_n_s16(int32x4_t __p0, int16x4_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 - __noswap_vmull_s16(__p1, (int16x4_t) {__p2, __p2, __p2, __p2});
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vset_lane_f16(__p0_260, __p1_260, __p2_260) __extension__ ({ \
-  float16_t __s0_260 = __p0_260; \
-  float16x4_t __s1_260 = __p1_260; \
-  float16x4_t __ret_260; \
-float16_t __reint_260 = __s0_260; \
-float16x4_t __reint1_260 = __s1_260; \
-int16x4_t __reint2_260 = vset_lane_s16(*(int16_t *) &__reint_260, *(int16x4_t *) &__reint1_260, __p2_260); \
-  __ret_260 = *(float16x4_t *) &__reint2_260; \
-  __ret_260; \
-})
-#else
-#define vset_lane_f16(__p0_261, __p1_261, __p2_261) __extension__ ({ \
-  float16_t __s0_261 = __p0_261; \
-  float16x4_t __s1_261 = __p1_261; \
-  float16x4_t __rev1_261;  __rev1_261 = __builtin_shufflevector(__s1_261, __s1_261, 3, 2, 1, 0); \
-  float16x4_t __ret_261; \
-float16_t __reint_261 = __s0_261; \
-float16x4_t __reint1_261 = __rev1_261; \
-int16x4_t __reint2_261 = __noswap_vset_lane_s16(*(int16_t *) &__reint_261, *(int16x4_t *) &__reint1_261, __p2_261); \
-  __ret_261 = *(float16x4_t *) &__reint2_261; \
-  __ret_261 = __builtin_shufflevector(__ret_261, __ret_261, 3, 2, 1, 0); \
-  __ret_261; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vsetq_lane_f16(__p0_262, __p1_262, __p2_262) __extension__ ({ \
-  float16_t __s0_262 = __p0_262; \
-  float16x8_t __s1_262 = __p1_262; \
-  float16x8_t __ret_262; \
-float16_t __reint_262 = __s0_262; \
-float16x8_t __reint1_262 = __s1_262; \
-int16x8_t __reint2_262 = vsetq_lane_s16(*(int16_t *) &__reint_262, *(int16x8_t *) &__reint1_262, __p2_262); \
-  __ret_262 = *(float16x8_t *) &__reint2_262; \
-  __ret_262; \
-})
-#else
-#define vsetq_lane_f16(__p0_263, __p1_263, __p2_263) __extension__ ({ \
-  float16_t __s0_263 = __p0_263; \
-  float16x8_t __s1_263 = __p1_263; \
-  float16x8_t __rev1_263;  __rev1_263 = __builtin_shufflevector(__s1_263, __s1_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __ret_263; \
-float16_t __reint_263 = __s0_263; \
-float16x8_t __reint1_263 = __rev1_263; \
-int16x8_t __reint2_263 = __noswap_vsetq_lane_s16(*(int16_t *) &__reint_263, *(int16x8_t *) &__reint1_263, __p2_263); \
-  __ret_263 = *(float16x8_t *) &__reint2_263; \
-  __ret_263 = __builtin_shufflevector(__ret_263, __ret_263, 7, 6, 5, 4, 3, 2, 1, 0); \
-  __ret_263; \
-})
-#endif
-
-#if defined(__ARM_FEATURE_FP16FML) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_high_f16(__p0_264, __p1_264, __p2_264, __p3_264) __extension__ ({ \
-  float32x4_t __s0_264 = __p0_264; \
-  float16x8_t __s1_264 = __p1_264; \
-  float16x4_t __s2_264 = __p2_264; \
-  float32x4_t __ret_264; \
-  __ret_264 = vfmlalq_high_f16(__s0_264, __s1_264, (float16x8_t) {vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264), vget_lane_f16(__s2_264, __p3_264)}); \
-  __ret_264; \
-})
-#else
-#define vfmlalq_lane_high_f16(__p0_265, __p1_265, __p2_265, __p3_265) __extension__ ({ \
-  float32x4_t __s0_265 = __p0_265; \
-  float16x8_t __s1_265 = __p1_265; \
-  float16x4_t __s2_265 = __p2_265; \
-  float32x4_t __rev0_265;  __rev0_265 = __builtin_shufflevector(__s0_265, __s0_265, 3, 2, 1, 0); \
-  float16x8_t __rev1_265;  __rev1_265 = __builtin_shufflevector(__s1_265, __s1_265, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_265;  __rev2_265 = __builtin_shufflevector(__s2_265, __s2_265, 3, 2, 1, 0); \
-  float32x4_t __ret_265; \
-  __ret_265 = __noswap_vfmlalq_high_f16(__rev0_265, __rev1_265, (float16x8_t) {__noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265), __noswap_vget_lane_f16(__rev2_265, __p3_265)}); \
-  __ret_265 = __builtin_shufflevector(__ret_265, __ret_265, 3, 2, 1, 0); \
-  __ret_265; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_high_f16(__p0_266, __p1_266, __p2_266, __p3_266) __extension__ ({ \
-  float32x2_t __s0_266 = __p0_266; \
-  float16x4_t __s1_266 = __p1_266; \
-  float16x4_t __s2_266 = __p2_266; \
-  float32x2_t __ret_266; \
-  __ret_266 = vfmlal_high_f16(__s0_266, __s1_266, (float16x4_t) {vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266), vget_lane_f16(__s2_266, __p3_266)}); \
-  __ret_266; \
-})
-#else
-#define vfmlal_lane_high_f16(__p0_267, __p1_267, __p2_267, __p3_267) __extension__ ({ \
-  float32x2_t __s0_267 = __p0_267; \
-  float16x4_t __s1_267 = __p1_267; \
-  float16x4_t __s2_267 = __p2_267; \
-  float32x2_t __rev0_267;  __rev0_267 = __builtin_shufflevector(__s0_267, __s0_267, 1, 0); \
-  float16x4_t __rev1_267;  __rev1_267 = __builtin_shufflevector(__s1_267, __s1_267, 3, 2, 1, 0); \
-  float16x4_t __rev2_267;  __rev2_267 = __builtin_shufflevector(__s2_267, __s2_267, 3, 2, 1, 0); \
-  float32x2_t __ret_267; \
-  __ret_267 = __noswap_vfmlal_high_f16(__rev0_267, __rev1_267, (float16x4_t) {__noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267), __noswap_vget_lane_f16(__rev2_267, __p3_267)}); \
-  __ret_267 = __builtin_shufflevector(__ret_267, __ret_267, 1, 0); \
-  __ret_267; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlalq_lane_low_f16(__p0_268, __p1_268, __p2_268, __p3_268) __extension__ ({ \
-  float32x4_t __s0_268 = __p0_268; \
-  float16x8_t __s1_268 = __p1_268; \
-  float16x4_t __s2_268 = __p2_268; \
-  float32x4_t __ret_268; \
-  __ret_268 = vfmlalq_low_f16(__s0_268, __s1_268, (float16x8_t) {vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268), vget_lane_f16(__s2_268, __p3_268)}); \
-  __ret_268; \
-})
-#else
-#define vfmlalq_lane_low_f16(__p0_269, __p1_269, __p2_269, __p3_269) __extension__ ({ \
-  float32x4_t __s0_269 = __p0_269; \
-  float16x8_t __s1_269 = __p1_269; \
-  float16x4_t __s2_269 = __p2_269; \
-  float32x4_t __rev0_269;  __rev0_269 = __builtin_shufflevector(__s0_269, __s0_269, 3, 2, 1, 0); \
-  float16x8_t __rev1_269;  __rev1_269 = __builtin_shufflevector(__s1_269, __s1_269, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_269;  __rev2_269 = __builtin_shufflevector(__s2_269, __s2_269, 3, 2, 1, 0); \
-  float32x4_t __ret_269; \
-  __ret_269 = __noswap_vfmlalq_low_f16(__rev0_269, __rev1_269, (float16x8_t) {__noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269), __noswap_vget_lane_f16(__rev2_269, __p3_269)}); \
-  __ret_269 = __builtin_shufflevector(__ret_269, __ret_269, 3, 2, 1, 0); \
-  __ret_269; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlal_lane_low_f16(__p0_270, __p1_270, __p2_270, __p3_270) __extension__ ({ \
-  float32x2_t __s0_270 = __p0_270; \
-  float16x4_t __s1_270 = __p1_270; \
-  float16x4_t __s2_270 = __p2_270; \
-  float32x2_t __ret_270; \
-  __ret_270 = vfmlal_low_f16(__s0_270, __s1_270, (float16x4_t) {vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270), vget_lane_f16(__s2_270, __p3_270)}); \
-  __ret_270; \
-})
-#else
-#define vfmlal_lane_low_f16(__p0_271, __p1_271, __p2_271, __p3_271) __extension__ ({ \
-  float32x2_t __s0_271 = __p0_271; \
-  float16x4_t __s1_271 = __p1_271; \
-  float16x4_t __s2_271 = __p2_271; \
-  float32x2_t __rev0_271;  __rev0_271 = __builtin_shufflevector(__s0_271, __s0_271, 1, 0); \
-  float16x4_t __rev1_271;  __rev1_271 = __builtin_shufflevector(__s1_271, __s1_271, 3, 2, 1, 0); \
-  float16x4_t __rev2_271;  __rev2_271 = __builtin_shufflevector(__s2_271, __s2_271, 3, 2, 1, 0); \
-  float32x2_t __ret_271; \
-  __ret_271 = __noswap_vfmlal_low_f16(__rev0_271, __rev1_271, (float16x4_t) {__noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271), __noswap_vget_lane_f16(__rev2_271, __p3_271)}); \
-  __ret_271 = __builtin_shufflevector(__ret_271, __ret_271, 1, 0); \
-  __ret_271; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_high_f16(__p0_272, __p1_272, __p2_272, __p3_272) __extension__ ({ \
-  float32x4_t __s0_272 = __p0_272; \
-  float16x8_t __s1_272 = __p1_272; \
-  float16x8_t __s2_272 = __p2_272; \
-  float32x4_t __ret_272; \
-  __ret_272 = vfmlalq_high_f16(__s0_272, __s1_272, (float16x8_t) {vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272), vgetq_lane_f16(__s2_272, __p3_272)}); \
-  __ret_272; \
-})
-#else
-#define vfmlalq_laneq_high_f16(__p0_273, __p1_273, __p2_273, __p3_273) __extension__ ({ \
-  float32x4_t __s0_273 = __p0_273; \
-  float16x8_t __s1_273 = __p1_273; \
-  float16x8_t __s2_273 = __p2_273; \
-  float32x4_t __rev0_273;  __rev0_273 = __builtin_shufflevector(__s0_273, __s0_273, 3, 2, 1, 0); \
-  float16x8_t __rev1_273;  __rev1_273 = __builtin_shufflevector(__s1_273, __s1_273, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_273;  __rev2_273 = __builtin_shufflevector(__s2_273, __s2_273, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_273; \
-  __ret_273 = __noswap_vfmlalq_high_f16(__rev0_273, __rev1_273, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273), __noswap_vgetq_lane_f16(__rev2_273, __p3_273)}); \
-  __ret_273 = __builtin_shufflevector(__ret_273, __ret_273, 3, 2, 1, 0); \
-  __ret_273; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_high_f16(__p0_274, __p1_274, __p2_274, __p3_274) __extension__ ({ \
-  float32x2_t __s0_274 = __p0_274; \
-  float16x4_t __s1_274 = __p1_274; \
-  float16x8_t __s2_274 = __p2_274; \
-  float32x2_t __ret_274; \
-  __ret_274 = vfmlal_high_f16(__s0_274, __s1_274, (float16x4_t) {vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274), vgetq_lane_f16(__s2_274, __p3_274)}); \
-  __ret_274; \
-})
-#else
-#define vfmlal_laneq_high_f16(__p0_275, __p1_275, __p2_275, __p3_275) __extension__ ({ \
-  float32x2_t __s0_275 = __p0_275; \
-  float16x4_t __s1_275 = __p1_275; \
-  float16x8_t __s2_275 = __p2_275; \
-  float32x2_t __rev0_275;  __rev0_275 = __builtin_shufflevector(__s0_275, __s0_275, 1, 0); \
-  float16x4_t __rev1_275;  __rev1_275 = __builtin_shufflevector(__s1_275, __s1_275, 3, 2, 1, 0); \
-  float16x8_t __rev2_275;  __rev2_275 = __builtin_shufflevector(__s2_275, __s2_275, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_275; \
-  __ret_275 = __noswap_vfmlal_high_f16(__rev0_275, __rev1_275, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275), __noswap_vgetq_lane_f16(__rev2_275, __p3_275)}); \
-  __ret_275 = __builtin_shufflevector(__ret_275, __ret_275, 1, 0); \
-  __ret_275; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlalq_laneq_low_f16(__p0_276, __p1_276, __p2_276, __p3_276) __extension__ ({ \
-  float32x4_t __s0_276 = __p0_276; \
-  float16x8_t __s1_276 = __p1_276; \
-  float16x8_t __s2_276 = __p2_276; \
-  float32x4_t __ret_276; \
-  __ret_276 = vfmlalq_low_f16(__s0_276, __s1_276, (float16x8_t) {vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276), vgetq_lane_f16(__s2_276, __p3_276)}); \
-  __ret_276; \
-})
-#else
-#define vfmlalq_laneq_low_f16(__p0_277, __p1_277, __p2_277, __p3_277) __extension__ ({ \
-  float32x4_t __s0_277 = __p0_277; \
-  float16x8_t __s1_277 = __p1_277; \
-  float16x8_t __s2_277 = __p2_277; \
-  float32x4_t __rev0_277;  __rev0_277 = __builtin_shufflevector(__s0_277, __s0_277, 3, 2, 1, 0); \
-  float16x8_t __rev1_277;  __rev1_277 = __builtin_shufflevector(__s1_277, __s1_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_277;  __rev2_277 = __builtin_shufflevector(__s2_277, __s2_277, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_277; \
-  __ret_277 = __noswap_vfmlalq_low_f16(__rev0_277, __rev1_277, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277), __noswap_vgetq_lane_f16(__rev2_277, __p3_277)}); \
-  __ret_277 = __builtin_shufflevector(__ret_277, __ret_277, 3, 2, 1, 0); \
-  __ret_277; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlal_laneq_low_f16(__p0_278, __p1_278, __p2_278, __p3_278) __extension__ ({ \
-  float32x2_t __s0_278 = __p0_278; \
-  float16x4_t __s1_278 = __p1_278; \
-  float16x8_t __s2_278 = __p2_278; \
-  float32x2_t __ret_278; \
-  __ret_278 = vfmlal_low_f16(__s0_278, __s1_278, (float16x4_t) {vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278), vgetq_lane_f16(__s2_278, __p3_278)}); \
-  __ret_278; \
-})
-#else
-#define vfmlal_laneq_low_f16(__p0_279, __p1_279, __p2_279, __p3_279) __extension__ ({ \
-  float32x2_t __s0_279 = __p0_279; \
-  float16x4_t __s1_279 = __p1_279; \
-  float16x8_t __s2_279 = __p2_279; \
-  float32x2_t __rev0_279;  __rev0_279 = __builtin_shufflevector(__s0_279, __s0_279, 1, 0); \
-  float16x4_t __rev1_279;  __rev1_279 = __builtin_shufflevector(__s1_279, __s1_279, 3, 2, 1, 0); \
-  float16x8_t __rev2_279;  __rev2_279 = __builtin_shufflevector(__s2_279, __s2_279, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_279; \
-  __ret_279 = __noswap_vfmlal_low_f16(__rev0_279, __rev1_279, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279), __noswap_vgetq_lane_f16(__rev2_279, __p3_279)}); \
-  __ret_279 = __builtin_shufflevector(__ret_279, __ret_279, 1, 0); \
-  __ret_279; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_high_f16(__p0_280, __p1_280, __p2_280, __p3_280) __extension__ ({ \
-  float32x4_t __s0_280 = __p0_280; \
-  float16x8_t __s1_280 = __p1_280; \
-  float16x4_t __s2_280 = __p2_280; \
-  float32x4_t __ret_280; \
-  __ret_280 = vfmlslq_high_f16(__s0_280, __s1_280, (float16x8_t) {vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280), vget_lane_f16(__s2_280, __p3_280)}); \
-  __ret_280; \
-})
-#else
-#define vfmlslq_lane_high_f16(__p0_281, __p1_281, __p2_281, __p3_281) __extension__ ({ \
-  float32x4_t __s0_281 = __p0_281; \
-  float16x8_t __s1_281 = __p1_281; \
-  float16x4_t __s2_281 = __p2_281; \
-  float32x4_t __rev0_281;  __rev0_281 = __builtin_shufflevector(__s0_281, __s0_281, 3, 2, 1, 0); \
-  float16x8_t __rev1_281;  __rev1_281 = __builtin_shufflevector(__s1_281, __s1_281, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_281;  __rev2_281 = __builtin_shufflevector(__s2_281, __s2_281, 3, 2, 1, 0); \
-  float32x4_t __ret_281; \
-  __ret_281 = __noswap_vfmlslq_high_f16(__rev0_281, __rev1_281, (float16x8_t) {__noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281), __noswap_vget_lane_f16(__rev2_281, __p3_281)}); \
-  __ret_281 = __builtin_shufflevector(__ret_281, __ret_281, 3, 2, 1, 0); \
-  __ret_281; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_high_f16(__p0_282, __p1_282, __p2_282, __p3_282) __extension__ ({ \
-  float32x2_t __s0_282 = __p0_282; \
-  float16x4_t __s1_282 = __p1_282; \
-  float16x4_t __s2_282 = __p2_282; \
-  float32x2_t __ret_282; \
-  __ret_282 = vfmlsl_high_f16(__s0_282, __s1_282, (float16x4_t) {vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282), vget_lane_f16(__s2_282, __p3_282)}); \
-  __ret_282; \
-})
-#else
-#define vfmlsl_lane_high_f16(__p0_283, __p1_283, __p2_283, __p3_283) __extension__ ({ \
-  float32x2_t __s0_283 = __p0_283; \
-  float16x4_t __s1_283 = __p1_283; \
-  float16x4_t __s2_283 = __p2_283; \
-  float32x2_t __rev0_283;  __rev0_283 = __builtin_shufflevector(__s0_283, __s0_283, 1, 0); \
-  float16x4_t __rev1_283;  __rev1_283 = __builtin_shufflevector(__s1_283, __s1_283, 3, 2, 1, 0); \
-  float16x4_t __rev2_283;  __rev2_283 = __builtin_shufflevector(__s2_283, __s2_283, 3, 2, 1, 0); \
-  float32x2_t __ret_283; \
-  __ret_283 = __noswap_vfmlsl_high_f16(__rev0_283, __rev1_283, (float16x4_t) {__noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283), __noswap_vget_lane_f16(__rev2_283, __p3_283)}); \
-  __ret_283 = __builtin_shufflevector(__ret_283, __ret_283, 1, 0); \
-  __ret_283; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_lane_low_f16(__p0_284, __p1_284, __p2_284, __p3_284) __extension__ ({ \
-  float32x4_t __s0_284 = __p0_284; \
-  float16x8_t __s1_284 = __p1_284; \
-  float16x4_t __s2_284 = __p2_284; \
-  float32x4_t __ret_284; \
-  __ret_284 = vfmlslq_low_f16(__s0_284, __s1_284, (float16x8_t) {vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284), vget_lane_f16(__s2_284, __p3_284)}); \
-  __ret_284; \
-})
-#else
-#define vfmlslq_lane_low_f16(__p0_285, __p1_285, __p2_285, __p3_285) __extension__ ({ \
-  float32x4_t __s0_285 = __p0_285; \
-  float16x8_t __s1_285 = __p1_285; \
-  float16x4_t __s2_285 = __p2_285; \
-  float32x4_t __rev0_285;  __rev0_285 = __builtin_shufflevector(__s0_285, __s0_285, 3, 2, 1, 0); \
-  float16x8_t __rev1_285;  __rev1_285 = __builtin_shufflevector(__s1_285, __s1_285, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x4_t __rev2_285;  __rev2_285 = __builtin_shufflevector(__s2_285, __s2_285, 3, 2, 1, 0); \
-  float32x4_t __ret_285; \
-  __ret_285 = __noswap_vfmlslq_low_f16(__rev0_285, __rev1_285, (float16x8_t) {__noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285), __noswap_vget_lane_f16(__rev2_285, __p3_285)}); \
-  __ret_285 = __builtin_shufflevector(__ret_285, __ret_285, 3, 2, 1, 0); \
-  __ret_285; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_lane_low_f16(__p0_286, __p1_286, __p2_286, __p3_286) __extension__ ({ \
-  float32x2_t __s0_286 = __p0_286; \
-  float16x4_t __s1_286 = __p1_286; \
-  float16x4_t __s2_286 = __p2_286; \
-  float32x2_t __ret_286; \
-  __ret_286 = vfmlsl_low_f16(__s0_286, __s1_286, (float16x4_t) {vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286), vget_lane_f16(__s2_286, __p3_286)}); \
-  __ret_286; \
-})
-#else
-#define vfmlsl_lane_low_f16(__p0_287, __p1_287, __p2_287, __p3_287) __extension__ ({ \
-  float32x2_t __s0_287 = __p0_287; \
-  float16x4_t __s1_287 = __p1_287; \
-  float16x4_t __s2_287 = __p2_287; \
-  float32x2_t __rev0_287;  __rev0_287 = __builtin_shufflevector(__s0_287, __s0_287, 1, 0); \
-  float16x4_t __rev1_287;  __rev1_287 = __builtin_shufflevector(__s1_287, __s1_287, 3, 2, 1, 0); \
-  float16x4_t __rev2_287;  __rev2_287 = __builtin_shufflevector(__s2_287, __s2_287, 3, 2, 1, 0); \
-  float32x2_t __ret_287; \
-  __ret_287 = __noswap_vfmlsl_low_f16(__rev0_287, __rev1_287, (float16x4_t) {__noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287), __noswap_vget_lane_f16(__rev2_287, __p3_287)}); \
-  __ret_287 = __builtin_shufflevector(__ret_287, __ret_287, 1, 0); \
-  __ret_287; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_high_f16(__p0_288, __p1_288, __p2_288, __p3_288) __extension__ ({ \
-  float32x4_t __s0_288 = __p0_288; \
-  float16x8_t __s1_288 = __p1_288; \
-  float16x8_t __s2_288 = __p2_288; \
-  float32x4_t __ret_288; \
-  __ret_288 = vfmlslq_high_f16(__s0_288, __s1_288, (float16x8_t) {vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288), vgetq_lane_f16(__s2_288, __p3_288)}); \
-  __ret_288; \
-})
-#else
-#define vfmlslq_laneq_high_f16(__p0_289, __p1_289, __p2_289, __p3_289) __extension__ ({ \
-  float32x4_t __s0_289 = __p0_289; \
-  float16x8_t __s1_289 = __p1_289; \
-  float16x8_t __s2_289 = __p2_289; \
-  float32x4_t __rev0_289;  __rev0_289 = __builtin_shufflevector(__s0_289, __s0_289, 3, 2, 1, 0); \
-  float16x8_t __rev1_289;  __rev1_289 = __builtin_shufflevector(__s1_289, __s1_289, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_289;  __rev2_289 = __builtin_shufflevector(__s2_289, __s2_289, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_289; \
-  __ret_289 = __noswap_vfmlslq_high_f16(__rev0_289, __rev1_289, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289), __noswap_vgetq_lane_f16(__rev2_289, __p3_289)}); \
-  __ret_289 = __builtin_shufflevector(__ret_289, __ret_289, 3, 2, 1, 0); \
-  __ret_289; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_high_f16(__p0_290, __p1_290, __p2_290, __p3_290) __extension__ ({ \
-  float32x2_t __s0_290 = __p0_290; \
-  float16x4_t __s1_290 = __p1_290; \
-  float16x8_t __s2_290 = __p2_290; \
-  float32x2_t __ret_290; \
-  __ret_290 = vfmlsl_high_f16(__s0_290, __s1_290, (float16x4_t) {vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290), vgetq_lane_f16(__s2_290, __p3_290)}); \
-  __ret_290; \
-})
-#else
-#define vfmlsl_laneq_high_f16(__p0_291, __p1_291, __p2_291, __p3_291) __extension__ ({ \
-  float32x2_t __s0_291 = __p0_291; \
-  float16x4_t __s1_291 = __p1_291; \
-  float16x8_t __s2_291 = __p2_291; \
-  float32x2_t __rev0_291;  __rev0_291 = __builtin_shufflevector(__s0_291, __s0_291, 1, 0); \
-  float16x4_t __rev1_291;  __rev1_291 = __builtin_shufflevector(__s1_291, __s1_291, 3, 2, 1, 0); \
-  float16x8_t __rev2_291;  __rev2_291 = __builtin_shufflevector(__s2_291, __s2_291, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_291; \
-  __ret_291 = __noswap_vfmlsl_high_f16(__rev0_291, __rev1_291, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291), __noswap_vgetq_lane_f16(__rev2_291, __p3_291)}); \
-  __ret_291 = __builtin_shufflevector(__ret_291, __ret_291, 1, 0); \
-  __ret_291; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlslq_laneq_low_f16(__p0_292, __p1_292, __p2_292, __p3_292) __extension__ ({ \
-  float32x4_t __s0_292 = __p0_292; \
-  float16x8_t __s1_292 = __p1_292; \
-  float16x8_t __s2_292 = __p2_292; \
-  float32x4_t __ret_292; \
-  __ret_292 = vfmlslq_low_f16(__s0_292, __s1_292, (float16x8_t) {vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292), vgetq_lane_f16(__s2_292, __p3_292)}); \
-  __ret_292; \
-})
-#else
-#define vfmlslq_laneq_low_f16(__p0_293, __p1_293, __p2_293, __p3_293) __extension__ ({ \
-  float32x4_t __s0_293 = __p0_293; \
-  float16x8_t __s1_293 = __p1_293; \
-  float16x8_t __s2_293 = __p2_293; \
-  float32x4_t __rev0_293;  __rev0_293 = __builtin_shufflevector(__s0_293, __s0_293, 3, 2, 1, 0); \
-  float16x8_t __rev1_293;  __rev1_293 = __builtin_shufflevector(__s1_293, __s1_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16x8_t __rev2_293;  __rev2_293 = __builtin_shufflevector(__s2_293, __s2_293, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x4_t __ret_293; \
-  __ret_293 = __noswap_vfmlslq_low_f16(__rev0_293, __rev1_293, (float16x8_t) {__noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293), __noswap_vgetq_lane_f16(__rev2_293, __p3_293)}); \
-  __ret_293 = __builtin_shufflevector(__ret_293, __ret_293, 3, 2, 1, 0); \
-  __ret_293; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vfmlsl_laneq_low_f16(__p0_294, __p1_294, __p2_294, __p3_294) __extension__ ({ \
-  float32x2_t __s0_294 = __p0_294; \
-  float16x4_t __s1_294 = __p1_294; \
-  float16x8_t __s2_294 = __p2_294; \
-  float32x2_t __ret_294; \
-  __ret_294 = vfmlsl_low_f16(__s0_294, __s1_294, (float16x4_t) {vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294), vgetq_lane_f16(__s2_294, __p3_294)}); \
-  __ret_294; \
-})
-#else
-#define vfmlsl_laneq_low_f16(__p0_295, __p1_295, __p2_295, __p3_295) __extension__ ({ \
-  float32x2_t __s0_295 = __p0_295; \
-  float16x4_t __s1_295 = __p1_295; \
-  float16x8_t __s2_295 = __p2_295; \
-  float32x2_t __rev0_295;  __rev0_295 = __builtin_shufflevector(__s0_295, __s0_295, 1, 0); \
-  float16x4_t __rev1_295;  __rev1_295 = __builtin_shufflevector(__s1_295, __s1_295, 3, 2, 1, 0); \
-  float16x8_t __rev2_295;  __rev2_295 = __builtin_shufflevector(__s2_295, __s2_295, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float32x2_t __ret_295; \
-  __ret_295 = __noswap_vfmlsl_low_f16(__rev0_295, __rev1_295, (float16x4_t) {__noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295), __noswap_vgetq_lane_f16(__rev2_295, __p3_295)}); \
-  __ret_295 = __builtin_shufflevector(__ret_295, __ret_295, 1, 0); \
-  __ret_295; \
-})
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-#define vmulh_lane_f16(__p0_296, __p1_296, __p2_296) __extension__ ({ \
-  float16_t __s0_296 = __p0_296; \
-  float16x4_t __s1_296 = __p1_296; \
-  float16_t __ret_296; \
-  __ret_296 = __s0_296 * vget_lane_f16(__s1_296, __p2_296); \
-  __ret_296; \
-})
-#else
-#define vmulh_lane_f16(__p0_297, __p1_297, __p2_297) __extension__ ({ \
-  float16_t __s0_297 = __p0_297; \
-  float16x4_t __s1_297 = __p1_297; \
-  float16x4_t __rev1_297;  __rev1_297 = __builtin_shufflevector(__s1_297, __s1_297, 3, 2, 1, 0); \
-  float16_t __ret_297; \
-  __ret_297 = __s0_297 * __noswap_vget_lane_f16(__rev1_297, __p2_297); \
-  __ret_297; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulh_laneq_f16(__p0_298, __p1_298, __p2_298) __extension__ ({ \
-  float16_t __s0_298 = __p0_298; \
-  float16x8_t __s1_298 = __p1_298; \
-  float16_t __ret_298; \
-  __ret_298 = __s0_298 * vgetq_lane_f16(__s1_298, __p2_298); \
-  __ret_298; \
-})
-#else
-#define vmulh_laneq_f16(__p0_299, __p1_299, __p2_299) __extension__ ({ \
-  float16_t __s0_299 = __p0_299; \
-  float16x8_t __s1_299 = __p1_299; \
-  float16x8_t __rev1_299;  __rev1_299 = __builtin_shufflevector(__s1_299, __s1_299, 7, 6, 5, 4, 3, 2, 1, 0); \
-  float16_t __ret_299; \
-  __ret_299 = __s0_299 * __noswap_vgetq_lane_f16(__rev1_299, __p2_299); \
-  __ret_299; \
-})
-#endif
-
-#endif
-#if defined(__ARM_FEATURE_QRDMX) && defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = vqadds_s32(__p0, vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32_t vqrdmlahs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqadds_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = vqaddh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16_t vqrdmlahh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqaddh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_lane_s32(__p0_300, __p1_300, __p2_300, __p3_300) __extension__ ({ \
-  int32_t __s0_300 = __p0_300; \
-  int32_t __s1_300 = __p1_300; \
-  int32x2_t __s2_300 = __p2_300; \
-  int32_t __ret_300; \
-  __ret_300 = vqadds_s32(__s0_300, vqrdmulhs_s32(__s1_300, vget_lane_s32(__s2_300, __p3_300))); \
-  __ret_300; \
-})
-#else
-#define vqrdmlahs_lane_s32(__p0_301, __p1_301, __p2_301, __p3_301) __extension__ ({ \
-  int32_t __s0_301 = __p0_301; \
-  int32_t __s1_301 = __p1_301; \
-  int32x2_t __s2_301 = __p2_301; \
-  int32x2_t __rev2_301;  __rev2_301 = __builtin_shufflevector(__s2_301, __s2_301, 1, 0); \
-  int32_t __ret_301; \
-  __ret_301 = __noswap_vqadds_s32(__s0_301, __noswap_vqrdmulhs_s32(__s1_301, __noswap_vget_lane_s32(__rev2_301, __p3_301))); \
-  __ret_301; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_lane_s16(__p0_302, __p1_302, __p2_302, __p3_302) __extension__ ({ \
-  int16_t __s0_302 = __p0_302; \
-  int16_t __s1_302 = __p1_302; \
-  int16x4_t __s2_302 = __p2_302; \
-  int16_t __ret_302; \
-  __ret_302 = vqaddh_s16(__s0_302, vqrdmulhh_s16(__s1_302, vget_lane_s16(__s2_302, __p3_302))); \
-  __ret_302; \
-})
-#else
-#define vqrdmlahh_lane_s16(__p0_303, __p1_303, __p2_303, __p3_303) __extension__ ({ \
-  int16_t __s0_303 = __p0_303; \
-  int16_t __s1_303 = __p1_303; \
-  int16x4_t __s2_303 = __p2_303; \
-  int16x4_t __rev2_303;  __rev2_303 = __builtin_shufflevector(__s2_303, __s2_303, 3, 2, 1, 0); \
-  int16_t __ret_303; \
-  __ret_303 = __noswap_vqaddh_s16(__s0_303, __noswap_vqrdmulhh_s16(__s1_303, __noswap_vget_lane_s16(__rev2_303, __p3_303))); \
-  __ret_303; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahs_laneq_s32(__p0_304, __p1_304, __p2_304, __p3_304) __extension__ ({ \
-  int32_t __s0_304 = __p0_304; \
-  int32_t __s1_304 = __p1_304; \
-  int32x4_t __s2_304 = __p2_304; \
-  int32_t __ret_304; \
-  __ret_304 = vqadds_s32(__s0_304, vqrdmulhs_s32(__s1_304, vgetq_lane_s32(__s2_304, __p3_304))); \
-  __ret_304; \
-})
-#else
-#define vqrdmlahs_laneq_s32(__p0_305, __p1_305, __p2_305, __p3_305) __extension__ ({ \
-  int32_t __s0_305 = __p0_305; \
-  int32_t __s1_305 = __p1_305; \
-  int32x4_t __s2_305 = __p2_305; \
-  int32x4_t __rev2_305;  __rev2_305 = __builtin_shufflevector(__s2_305, __s2_305, 3, 2, 1, 0); \
-  int32_t __ret_305; \
-  __ret_305 = __noswap_vqadds_s32(__s0_305, __noswap_vqrdmulhs_s32(__s1_305, __noswap_vgetq_lane_s32(__rev2_305, __p3_305))); \
-  __ret_305; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlahh_laneq_s16(__p0_306, __p1_306, __p2_306, __p3_306) __extension__ ({ \
-  int16_t __s0_306 = __p0_306; \
-  int16_t __s1_306 = __p1_306; \
-  int16x8_t __s2_306 = __p2_306; \
-  int16_t __ret_306; \
-  __ret_306 = vqaddh_s16(__s0_306, vqrdmulhh_s16(__s1_306, vgetq_lane_s16(__s2_306, __p3_306))); \
-  __ret_306; \
-})
-#else
-#define vqrdmlahh_laneq_s16(__p0_307, __p1_307, __p2_307, __p3_307) __extension__ ({ \
-  int16_t __s0_307 = __p0_307; \
-  int16_t __s1_307 = __p1_307; \
-  int16x8_t __s2_307 = __p2_307; \
-  int16x8_t __rev2_307;  __rev2_307 = __builtin_shufflevector(__s2_307, __s2_307, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_307; \
-  __ret_307 = __noswap_vqaddh_s16(__s0_307, __noswap_vqrdmulhh_s16(__s1_307, __noswap_vgetq_lane_s16(__rev2_307, __p3_307))); \
-  __ret_307; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = vqsubs_s32(__p0, vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int32_t vqrdmlshs_s32(int32_t __p0, int32_t __p1, int32_t __p2) {
-  int32_t __ret;
-  __ret = __noswap_vqsubs_s32(__p0, __noswap_vqrdmulhs_s32(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = vqsubh_s16(__p0, vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
-#else
-__ai int16_t vqrdmlshh_s16(int16_t __p0, int16_t __p1, int16_t __p2) {
-  int16_t __ret;
-  __ret = __noswap_vqsubh_s16(__p0, __noswap_vqrdmulhh_s16(__p1, __p2));
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_lane_s32(__p0_308, __p1_308, __p2_308, __p3_308) __extension__ ({ \
-  int32_t __s0_308 = __p0_308; \
-  int32_t __s1_308 = __p1_308; \
-  int32x2_t __s2_308 = __p2_308; \
-  int32_t __ret_308; \
-  __ret_308 = vqsubs_s32(__s0_308, vqrdmulhs_s32(__s1_308, vget_lane_s32(__s2_308, __p3_308))); \
-  __ret_308; \
-})
-#else
-#define vqrdmlshs_lane_s32(__p0_309, __p1_309, __p2_309, __p3_309) __extension__ ({ \
-  int32_t __s0_309 = __p0_309; \
-  int32_t __s1_309 = __p1_309; \
-  int32x2_t __s2_309 = __p2_309; \
-  int32x2_t __rev2_309;  __rev2_309 = __builtin_shufflevector(__s2_309, __s2_309, 1, 0); \
-  int32_t __ret_309; \
-  __ret_309 = __noswap_vqsubs_s32(__s0_309, __noswap_vqrdmulhs_s32(__s1_309, __noswap_vget_lane_s32(__rev2_309, __p3_309))); \
-  __ret_309; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_lane_s16(__p0_310, __p1_310, __p2_310, __p3_310) __extension__ ({ \
-  int16_t __s0_310 = __p0_310; \
-  int16_t __s1_310 = __p1_310; \
-  int16x4_t __s2_310 = __p2_310; \
-  int16_t __ret_310; \
-  __ret_310 = vqsubh_s16(__s0_310, vqrdmulhh_s16(__s1_310, vget_lane_s16(__s2_310, __p3_310))); \
-  __ret_310; \
-})
-#else
-#define vqrdmlshh_lane_s16(__p0_311, __p1_311, __p2_311, __p3_311) __extension__ ({ \
-  int16_t __s0_311 = __p0_311; \
-  int16_t __s1_311 = __p1_311; \
-  int16x4_t __s2_311 = __p2_311; \
-  int16x4_t __rev2_311;  __rev2_311 = __builtin_shufflevector(__s2_311, __s2_311, 3, 2, 1, 0); \
-  int16_t __ret_311; \
-  __ret_311 = __noswap_vqsubh_s16(__s0_311, __noswap_vqrdmulhh_s16(__s1_311, __noswap_vget_lane_s16(__rev2_311, __p3_311))); \
-  __ret_311; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshs_laneq_s32(__p0_312, __p1_312, __p2_312, __p3_312) __extension__ ({ \
-  int32_t __s0_312 = __p0_312; \
-  int32_t __s1_312 = __p1_312; \
-  int32x4_t __s2_312 = __p2_312; \
-  int32_t __ret_312; \
-  __ret_312 = vqsubs_s32(__s0_312, vqrdmulhs_s32(__s1_312, vgetq_lane_s32(__s2_312, __p3_312))); \
-  __ret_312; \
-})
-#else
-#define vqrdmlshs_laneq_s32(__p0_313, __p1_313, __p2_313, __p3_313) __extension__ ({ \
-  int32_t __s0_313 = __p0_313; \
-  int32_t __s1_313 = __p1_313; \
-  int32x4_t __s2_313 = __p2_313; \
-  int32x4_t __rev2_313;  __rev2_313 = __builtin_shufflevector(__s2_313, __s2_313, 3, 2, 1, 0); \
-  int32_t __ret_313; \
-  __ret_313 = __noswap_vqsubs_s32(__s0_313, __noswap_vqrdmulhs_s32(__s1_313, __noswap_vgetq_lane_s32(__rev2_313, __p3_313))); \
-  __ret_313; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vqrdmlshh_laneq_s16(__p0_314, __p1_314, __p2_314, __p3_314) __extension__ ({ \
-  int16_t __s0_314 = __p0_314; \
-  int16_t __s1_314 = __p1_314; \
-  int16x8_t __s2_314 = __p2_314; \
-  int16_t __ret_314; \
-  __ret_314 = vqsubh_s16(__s0_314, vqrdmulhh_s16(__s1_314, vgetq_lane_s16(__s2_314, __p3_314))); \
-  __ret_314; \
-})
-#else
-#define vqrdmlshh_laneq_s16(__p0_315, __p1_315, __p2_315, __p3_315) __extension__ ({ \
-  int16_t __s0_315 = __p0_315; \
-  int16_t __s1_315 = __p1_315; \
-  int16x8_t __s2_315 = __p2_315; \
-  int16x8_t __rev2_315;  __rev2_315 = __builtin_shufflevector(__s2_315, __s2_315, 7, 6, 5, 4, 3, 2, 1, 0); \
-  int16_t __ret_315; \
-  __ret_315 = __noswap_vqsubh_s16(__s0_315, __noswap_vqrdmulhh_s16(__s1_315, __noswap_vgetq_lane_s16(__rev2_315, __p3_315))); \
-  __ret_315; \
-})
-#endif
-
-#endif
-#if defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = vabdl_u8(vget_high_u8(__p0), vget_high_u8(__p1));
-  return __ret;
-}
-#else
-__ai uint16x8_t vabdl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vabdl_u8(__noswap_vget_high_u8(__rev0), __noswap_vget_high_u8(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = vabdl_u32(vget_high_u32(__p0), vget_high_u32(__p1));
-  return __ret;
-}
-#else
-__ai uint64x2_t vabdl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vabdl_u32(__noswap_vget_high_u32(__rev0), __noswap_vget_high_u32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = vabdl_u16(vget_high_u16(__p0), vget_high_u16(__p1));
-  return __ret;
-}
-#else
-__ai uint32x4_t vabdl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vabdl_u16(__noswap_vget_high_u16(__rev0), __noswap_vget_high_u16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = vabdl_s8(vget_high_s8(__p0), vget_high_s8(__p1));
-  return __ret;
-}
-#else
-__ai int16x8_t vabdl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vabdl_s8(__noswap_vget_high_s8(__rev0), __noswap_vget_high_s8(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = vabdl_s32(vget_high_s32(__p0), vget_high_s32(__p1));
-  return __ret;
-}
-#else
-__ai int64x2_t vabdl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vabdl_s32(__noswap_vget_high_s32(__rev0), __noswap_vget_high_s32(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = vabdl_s16(vget_high_s16(__p0), vget_high_s16(__p1));
-  return __ret;
-}
-#else
-__ai int32x4_t vabdl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vabdl_s16(__noswap_vget_high_s16(__rev0), __noswap_vget_high_s16(__rev1));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = vmovl_high_u8(__p0) + vmovl_high_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddl_high_u8(uint8x16_t __p0, uint8x16_t __p1) {
-  uint8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmovl_high_u8(__rev0) + __noswap_vmovl_high_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = vmovl_high_u32(__p0) + vmovl_high_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vaddl_high_u32(uint32x4_t __p0, uint32x4_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmovl_high_u32(__rev0) + __noswap_vmovl_high_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = vmovl_high_u16(__p0) + vmovl_high_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddl_high_u16(uint16x8_t __p0, uint16x8_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmovl_high_u16(__rev0) + __noswap_vmovl_high_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = vmovl_high_s8(__p0) + vmovl_high_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vaddl_high_s8(int8x16_t __p0, int8x16_t __p1) {
-  int8x16_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmovl_high_s8(__rev0) + __noswap_vmovl_high_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = vmovl_high_s32(__p0) + vmovl_high_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vaddl_high_s32(int32x4_t __p0, int32x4_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmovl_high_s32(__rev0) + __noswap_vmovl_high_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = vmovl_high_s16(__p0) + vmovl_high_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vaddl_high_s16(int16x8_t __p0, int16x8_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmovl_high_s16(__rev0) + __noswap_vmovl_high_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __ret;
-  __ret = __p0 + vmovl_high_u8(__p1);
-  return __ret;
-}
-#else
-__ai uint16x8_t vaddw_high_u8(uint16x8_t __p0, uint8x16_t __p1) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_u8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __ret;
-  __ret = __p0 + vmovl_high_u32(__p1);
-  return __ret;
-}
-#else
-__ai uint64x2_t vaddw_high_u32(uint64x2_t __p0, uint32x4_t __p1) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_u32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __ret;
-  __ret = __p0 + vmovl_high_u16(__p1);
-  return __ret;
-}
-#else
-__ai uint32x4_t vaddw_high_u16(uint32x4_t __p0, uint16x8_t __p1) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_u16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __ret;
-  __ret = __p0 + vmovl_high_s8(__p1);
-  return __ret;
-}
-#else
-__ai int16x8_t vaddw_high_s8(int16x8_t __p0, int8x16_t __p1) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_s8(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __ret;
-  __ret = __p0 + vmovl_high_s32(__p1);
-  return __ret;
-}
-#else
-__ai int64x2_t vaddw_high_s32(int64x2_t __p0, int32x4_t __p1) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_s32(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __ret;
-  __ret = __p0 + vmovl_high_s16(__p1);
-  return __ret;
-}
-#else
-__ai int32x4_t vaddw_high_s16(int32x4_t __p0, int16x8_t __p1) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vmovl_high_s16(__rev1);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_p64(__p0_316, __p1_316, __p2_316, __p3_316) __extension__ ({ \
-  poly64x2_t __s0_316 = __p0_316; \
-  poly64x1_t __s2_316 = __p2_316; \
-  poly64x2_t __ret_316; \
-  __ret_316 = vsetq_lane_p64(vget_lane_p64(__s2_316, __p3_316), __s0_316, __p1_316); \
-  __ret_316; \
-})
-#else
-#define vcopyq_lane_p64(__p0_317, __p1_317, __p2_317, __p3_317) __extension__ ({ \
-  poly64x2_t __s0_317 = __p0_317; \
-  poly64x1_t __s2_317 = __p2_317; \
-  poly64x2_t __rev0_317;  __rev0_317 = __builtin_shufflevector(__s0_317, __s0_317, 1, 0); \
-  poly64x2_t __ret_317; \
-  __ret_317 = __noswap_vsetq_lane_p64(__noswap_vget_lane_p64(__s2_317, __p3_317), __rev0_317, __p1_317); \
-  __ret_317 = __builtin_shufflevector(__ret_317, __ret_317, 1, 0); \
-  __ret_317; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_lane_f64(__p0_318, __p1_318, __p2_318, __p3_318) __extension__ ({ \
-  float64x2_t __s0_318 = __p0_318; \
-  float64x1_t __s2_318 = __p2_318; \
-  float64x2_t __ret_318; \
-  __ret_318 = vsetq_lane_f64(vget_lane_f64(__s2_318, __p3_318), __s0_318, __p1_318); \
-  __ret_318; \
-})
-#else
-#define vcopyq_lane_f64(__p0_319, __p1_319, __p2_319, __p3_319) __extension__ ({ \
-  float64x2_t __s0_319 = __p0_319; \
-  float64x1_t __s2_319 = __p2_319; \
-  float64x2_t __rev0_319;  __rev0_319 = __builtin_shufflevector(__s0_319, __s0_319, 1, 0); \
-  float64x2_t __ret_319; \
-  __ret_319 = __noswap_vsetq_lane_f64(__noswap_vget_lane_f64(__s2_319, __p3_319), __rev0_319, __p1_319); \
-  __ret_319 = __builtin_shufflevector(__ret_319, __ret_319, 1, 0); \
-  __ret_319; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_p64(__p0_320, __p1_320, __p2_320, __p3_320) __extension__ ({ \
-  poly64x1_t __s0_320 = __p0_320; \
-  poly64x1_t __s2_320 = __p2_320; \
-  poly64x1_t __ret_320; \
-  __ret_320 = vset_lane_p64(vget_lane_p64(__s2_320, __p3_320), __s0_320, __p1_320); \
-  __ret_320; \
-})
-#else
-#define vcopy_lane_p64(__p0_321, __p1_321, __p2_321, __p3_321) __extension__ ({ \
-  poly64x1_t __s0_321 = __p0_321; \
-  poly64x1_t __s2_321 = __p2_321; \
-  poly64x1_t __ret_321; \
-  __ret_321 = __noswap_vset_lane_p64(__noswap_vget_lane_p64(__s2_321, __p3_321), __s0_321, __p1_321); \
-  __ret_321; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_lane_f64(__p0_322, __p1_322, __p2_322, __p3_322) __extension__ ({ \
-  float64x1_t __s0_322 = __p0_322; \
-  float64x1_t __s2_322 = __p2_322; \
-  float64x1_t __ret_322; \
-  __ret_322 = vset_lane_f64(vget_lane_f64(__s2_322, __p3_322), __s0_322, __p1_322); \
-  __ret_322; \
-})
-#else
-#define vcopy_lane_f64(__p0_323, __p1_323, __p2_323, __p3_323) __extension__ ({ \
-  float64x1_t __s0_323 = __p0_323; \
-  float64x1_t __s2_323 = __p2_323; \
-  float64x1_t __ret_323; \
-  __ret_323 = __noswap_vset_lane_f64(__noswap_vget_lane_f64(__s2_323, __p3_323), __s0_323, __p1_323); \
-  __ret_323; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_p64(__p0_324, __p1_324, __p2_324, __p3_324) __extension__ ({ \
-  poly64x2_t __s0_324 = __p0_324; \
-  poly64x2_t __s2_324 = __p2_324; \
-  poly64x2_t __ret_324; \
-  __ret_324 = vsetq_lane_p64(vgetq_lane_p64(__s2_324, __p3_324), __s0_324, __p1_324); \
-  __ret_324; \
-})
-#else
-#define vcopyq_laneq_p64(__p0_325, __p1_325, __p2_325, __p3_325) __extension__ ({ \
-  poly64x2_t __s0_325 = __p0_325; \
-  poly64x2_t __s2_325 = __p2_325; \
-  poly64x2_t __rev0_325;  __rev0_325 = __builtin_shufflevector(__s0_325, __s0_325, 1, 0); \
-  poly64x2_t __rev2_325;  __rev2_325 = __builtin_shufflevector(__s2_325, __s2_325, 1, 0); \
-  poly64x2_t __ret_325; \
-  __ret_325 = __noswap_vsetq_lane_p64(__noswap_vgetq_lane_p64(__rev2_325, __p3_325), __rev0_325, __p1_325); \
-  __ret_325 = __builtin_shufflevector(__ret_325, __ret_325, 1, 0); \
-  __ret_325; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopyq_laneq_f64(__p0_326, __p1_326, __p2_326, __p3_326) __extension__ ({ \
-  float64x2_t __s0_326 = __p0_326; \
-  float64x2_t __s2_326 = __p2_326; \
-  float64x2_t __ret_326; \
-  __ret_326 = vsetq_lane_f64(vgetq_lane_f64(__s2_326, __p3_326), __s0_326, __p1_326); \
-  __ret_326; \
-})
-#else
-#define vcopyq_laneq_f64(__p0_327, __p1_327, __p2_327, __p3_327) __extension__ ({ \
-  float64x2_t __s0_327 = __p0_327; \
-  float64x2_t __s2_327 = __p2_327; \
-  float64x2_t __rev0_327;  __rev0_327 = __builtin_shufflevector(__s0_327, __s0_327, 1, 0); \
-  float64x2_t __rev2_327;  __rev2_327 = __builtin_shufflevector(__s2_327, __s2_327, 1, 0); \
-  float64x2_t __ret_327; \
-  __ret_327 = __noswap_vsetq_lane_f64(__noswap_vgetq_lane_f64(__rev2_327, __p3_327), __rev0_327, __p1_327); \
-  __ret_327 = __builtin_shufflevector(__ret_327, __ret_327, 1, 0); \
-  __ret_327; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_p64(__p0_328, __p1_328, __p2_328, __p3_328) __extension__ ({ \
-  poly64x1_t __s0_328 = __p0_328; \
-  poly64x2_t __s2_328 = __p2_328; \
-  poly64x1_t __ret_328; \
-  __ret_328 = vset_lane_p64(vgetq_lane_p64(__s2_328, __p3_328), __s0_328, __p1_328); \
-  __ret_328; \
-})
-#else
-#define vcopy_laneq_p64(__p0_329, __p1_329, __p2_329, __p3_329) __extension__ ({ \
-  poly64x1_t __s0_329 = __p0_329; \
-  poly64x2_t __s2_329 = __p2_329; \
-  poly64x2_t __rev2_329;  __rev2_329 = __builtin_shufflevector(__s2_329, __s2_329, 1, 0); \
-  poly64x1_t __ret_329; \
-  __ret_329 = __noswap_vset_lane_p64(__noswap_vgetq_lane_p64(__rev2_329, __p3_329), __s0_329, __p1_329); \
-  __ret_329; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vcopy_laneq_f64(__p0_330, __p1_330, __p2_330, __p3_330) __extension__ ({ \
-  float64x1_t __s0_330 = __p0_330; \
-  float64x2_t __s2_330 = __p2_330; \
-  float64x1_t __ret_330; \
-  __ret_330 = vset_lane_f64(vgetq_lane_f64(__s2_330, __p3_330), __s0_330, __p1_330); \
-  __ret_330; \
-})
-#else
-#define vcopy_laneq_f64(__p0_331, __p1_331, __p2_331, __p3_331) __extension__ ({ \
-  float64x1_t __s0_331 = __p0_331; \
-  float64x2_t __s2_331 = __p2_331; \
-  float64x2_t __rev2_331;  __rev2_331 = __builtin_shufflevector(__s2_331, __s2_331, 1, 0); \
-  float64x1_t __ret_331; \
-  __ret_331 = __noswap_vset_lane_f64(__noswap_vgetq_lane_f64(__rev2_331, __p3_331), __s0_331, __p1_331); \
-  __ret_331; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __ret;
-  __ret = vmlal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmlal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __ret;
-  __ret = vmlal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmlal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __ret;
-  __ret = vmlal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmlal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __ret;
-  __ret = vmlal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vmlal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmlal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __ret;
-  __ret = vmlal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
-  return __ret;
-}
-#else
-__ai int64x2_t vmlal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmlal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __ret;
-  __ret = vmlal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vmlal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmlal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = vmlal_n_u32(__p0, vget_high_u32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlal_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmlal_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = vmlal_n_u16(__p0, vget_high_u16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlal_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmlal_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = vmlal_n_s32(__p0, vget_high_s32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vmlal_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmlal_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = vmlal_n_s16(__p0, vget_high_s16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vmlal_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmlal_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __ret;
-  __ret = vmlsl_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vmlsl_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vmlsl_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __ret;
-  __ret = vmlsl_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlsl_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmlsl_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __ret;
-  __ret = vmlsl_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsl_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmlsl_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __ret;
-  __ret = vmlsl_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vmlsl_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vmlsl_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __ret;
-  __ret = vmlsl_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
-  return __ret;
-}
-#else
-__ai int64x2_t vmlsl_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmlsl_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __ret;
-  __ret = vmlsl_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsl_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmlsl_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint64x2_t __ret;
-  __ret = vmlsl_n_u32(__p0, vget_high_u32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vmlsl_high_n_u32(uint64x2_t __p0, uint32x4_t __p1, uint32_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vmlsl_n_u32(__rev0, __noswap_vget_high_u32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint32x4_t __ret;
-  __ret = vmlsl_n_u16(__p0, vget_high_u16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vmlsl_high_n_u16(uint32x4_t __p0, uint16x8_t __p1, uint16_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vmlsl_n_u16(__rev0, __noswap_vget_high_u16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __ret;
-  __ret = vmlsl_n_s32(__p0, vget_high_s32(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vmlsl_high_n_s32(int64x2_t __p0, int32x4_t __p1, int32_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vmlsl_n_s32(__rev0, __noswap_vget_high_s32(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __ret;
-  __ret = vmlsl_n_s16(__p0, vget_high_s16(__p1), __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vmlsl_high_n_s16(int32x4_t __p0, int16x8_t __p1, int16_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vmlsl_n_s16(__rev0, __noswap_vget_high_s16(__rev1), __p2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_lane_f64(__p0_332, __p1_332, __p2_332) __extension__ ({ \
-  float64x1_t __s0_332 = __p0_332; \
-  float64x1_t __s1_332 = __p1_332; \
-  float64x1_t __ret_332; \
-  float64_t __x_332 = vget_lane_f64(__s0_332, 0); \
-  float64_t __y_332 = vget_lane_f64(__s1_332, __p2_332); \
-  float64_t __z_332 = vmulxd_f64(__x_332, __y_332); \
-  __ret_332 = vset_lane_f64(__z_332, __s0_332, __p2_332); \
-  __ret_332; \
-})
-#else
-#define vmulx_lane_f64(__p0_333, __p1_333, __p2_333) __extension__ ({ \
-  float64x1_t __s0_333 = __p0_333; \
-  float64x1_t __s1_333 = __p1_333; \
-  float64x1_t __ret_333; \
-  float64_t __x_333 = __noswap_vget_lane_f64(__s0_333, 0); \
-  float64_t __y_333 = __noswap_vget_lane_f64(__s1_333, __p2_333); \
-  float64_t __z_333 = __noswap_vmulxd_f64(__x_333, __y_333); \
-  __ret_333 = __noswap_vset_lane_f64(__z_333, __s0_333, __p2_333); \
-  __ret_333; \
-})
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-#define vmulx_laneq_f64(__p0_334, __p1_334, __p2_334) __extension__ ({ \
-  float64x1_t __s0_334 = __p0_334; \
-  float64x2_t __s1_334 = __p1_334; \
-  float64x1_t __ret_334; \
-  float64_t __x_334 = vget_lane_f64(__s0_334, 0); \
-  float64_t __y_334 = vgetq_lane_f64(__s1_334, __p2_334); \
-  float64_t __z_334 = vmulxd_f64(__x_334, __y_334); \
-  __ret_334 = vset_lane_f64(__z_334, __s0_334, 0); \
-  __ret_334; \
-})
-#else
-#define vmulx_laneq_f64(__p0_335, __p1_335, __p2_335) __extension__ ({ \
-  float64x1_t __s0_335 = __p0_335; \
-  float64x2_t __s1_335 = __p1_335; \
-  float64x2_t __rev1_335;  __rev1_335 = __builtin_shufflevector(__s1_335, __s1_335, 1, 0); \
-  float64x1_t __ret_335; \
-  float64_t __x_335 = __noswap_vget_lane_f64(__s0_335, 0); \
-  float64_t __y_335 = __noswap_vgetq_lane_f64(__rev1_335, __p2_335); \
-  float64_t __z_335 = __noswap_vmulxd_f64(__x_335, __y_335); \
-  __ret_335 = __noswap_vset_lane_f64(__z_335, __s0_335, 0); \
-  __ret_335; \
-})
-#endif
-
-#endif
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + vabdl_u8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint16x8_t vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __rev0 + __noswap_vabdl_u8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint16x8_t __noswap_vabal_u8(uint16x8_t __p0, uint8x8_t __p1, uint8x8_t __p2) {
-  uint16x8_t __ret;
-  __ret = __p0 + __noswap_vabdl_u8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + vabdl_u32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint64x2_t vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  uint32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __rev0 + __noswap_vabdl_u32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai uint64x2_t __noswap_vabal_u32(uint64x2_t __p0, uint32x2_t __p1, uint32x2_t __p2) {
-  uint64x2_t __ret;
-  __ret = __p0 + __noswap_vabdl_u32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + vabdl_u16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai uint32x4_t vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __rev0 + __noswap_vabdl_u16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai uint32x4_t __noswap_vabal_u16(uint32x4_t __p0, uint16x4_t __p1, uint16x4_t __p2) {
-  uint32x4_t __ret;
-  __ret = __p0 + __noswap_vabdl_u16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + vabdl_s8(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int16x8_t vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __rev0 + __noswap_vabdl_s8(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int16x8_t __noswap_vabal_s8(int16x8_t __p0, int8x8_t __p1, int8x8_t __p2) {
-  int16x8_t __ret;
-  __ret = __p0 + __noswap_vabdl_s8(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + vabdl_s32(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int64x2_t vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x2_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 1, 0);
-  int32x2_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 1, 0);
-  int64x2_t __ret;
-  __ret = __rev0 + __noswap_vabdl_s32(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-__ai int64x2_t __noswap_vabal_s32(int64x2_t __p0, int32x2_t __p1, int32x2_t __p2) {
-  int64x2_t __ret;
-  __ret = __p0 + __noswap_vabdl_s32(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + vabdl_s16(__p1, __p2);
-  return __ret;
-}
-#else
-__ai int32x4_t vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int16x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __rev0 + __noswap_vabdl_s16(__rev1, __rev2);
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-__ai int32x4_t __noswap_vabal_s16(int32x4_t __p0, int16x4_t __p1, int16x4_t __p2) {
-  int32x4_t __ret;
-  __ret = __p0 + __noswap_vabdl_s16(__p1, __p2);
-  return __ret;
-}
-#endif
-
-#if defined(__aarch64__)
-#ifdef __LITTLE_ENDIAN__
-__ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __ret;
-  __ret = vabal_u8(__p0, vget_high_u8(__p1), vget_high_u8(__p2));
-  return __ret;
-}
-#else
-__ai uint16x8_t vabal_high_u8(uint16x8_t __p0, uint8x16_t __p1, uint8x16_t __p2) {
-  uint16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __ret;
-  __ret = __noswap_vabal_u8(__rev0, __noswap_vget_high_u8(__rev1), __noswap_vget_high_u8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __ret;
-  __ret = vabal_u32(__p0, vget_high_u32(__p1), vget_high_u32(__p2));
-  return __ret;
-}
-#else
-__ai uint64x2_t vabal_high_u32(uint64x2_t __p0, uint32x4_t __p1, uint32x4_t __p2) {
-  uint64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  uint32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  uint32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  uint64x2_t __ret;
-  __ret = __noswap_vabal_u32(__rev0, __noswap_vget_high_u32(__rev1), __noswap_vget_high_u32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __ret;
-  __ret = vabal_u16(__p0, vget_high_u16(__p1), vget_high_u16(__p2));
-  return __ret;
-}
-#else
-__ai uint32x4_t vabal_high_u16(uint32x4_t __p0, uint16x8_t __p1, uint16x8_t __p2) {
-  uint32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  uint16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  uint32x4_t __ret;
-  __ret = __noswap_vabal_u16(__rev0, __noswap_vget_high_u16(__rev1), __noswap_vget_high_u16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __ret;
-  __ret = vabal_s8(__p0, vget_high_s8(__p1), vget_high_s8(__p2));
-  return __ret;
-}
-#else
-__ai int16x8_t vabal_high_s8(int16x8_t __p0, int8x16_t __p1, int8x16_t __p2) {
-  int16x8_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int8x16_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __ret;
-  __ret = __noswap_vabal_s8(__rev0, __noswap_vget_high_s8(__rev1), __noswap_vget_high_s8(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 7, 6, 5, 4, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __ret;
-  __ret = vabal_s32(__p0, vget_high_s32(__p1), vget_high_s32(__p2));
-  return __ret;
-}
-#else
-__ai int64x2_t vabal_high_s32(int64x2_t __p0, int32x4_t __p1, int32x4_t __p2) {
-  int64x2_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 1, 0);
-  int32x4_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 3, 2, 1, 0);
-  int32x4_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 3, 2, 1, 0);
-  int64x2_t __ret;
-  __ret = __noswap_vabal_s32(__rev0, __noswap_vget_high_s32(__rev1), __noswap_vget_high_s32(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 1, 0);
-  return __ret;
-}
-#endif
-
-#ifdef __LITTLE_ENDIAN__
-__ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __ret;
-  __ret = vabal_s16(__p0, vget_high_s16(__p1), vget_high_s16(__p2));
-  return __ret;
-}
-#else
-__ai int32x4_t vabal_high_s16(int32x4_t __p0, int16x8_t __p1, int16x8_t __p2) {
-  int32x4_t __rev0;  __rev0 = __builtin_shufflevector(__p0, __p0, 3, 2, 1, 0);
-  int16x8_t __rev1;  __rev1 = __builtin_shufflevector(__p1, __p1, 7, 6, 5, 4, 3, 2, 1, 0);
-  int16x8_t __rev2;  __rev2 = __builtin_shufflevector(__p2, __p2, 7, 6, 5, 4, 3, 2, 1, 0);
-  int32x4_t __ret;
-  __ret = __noswap_vabal_s16(__rev0, __noswap_vget_high_s16(__rev1), __noswap_vget_high_s16(__rev2));
-  __ret = __builtin_shufflevector(__ret, __ret, 3, 2, 1, 0);
-  return __ret;
-}
-#endif
-
-#endif
-
-#undef __ai
-
-#endif /* __ARM_NEON_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/avx512fintrin.h b/linux-x86/lib64/clang/9.0.7/include/avx512fintrin.h
deleted file mode 100644
index 132761f..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/avx512fintrin.h
+++ /dev/null
@@ -1,9666 +0,0 @@
-/*===---- avx512fintrin.h - AVX512F intrinsics -----------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-#ifndef __IMMINTRIN_H
-#error "Never use <avx512fintrin.h> directly; include <immintrin.h> instead."
-#endif
-
-#ifndef __AVX512FINTRIN_H
-#define __AVX512FINTRIN_H
-
-typedef char __v64qi __attribute__((__vector_size__(64)));
-typedef short __v32hi __attribute__((__vector_size__(64)));
-typedef double __v8df __attribute__((__vector_size__(64)));
-typedef float __v16sf __attribute__((__vector_size__(64)));
-typedef long long __v8di __attribute__((__vector_size__(64)));
-typedef int __v16si __attribute__((__vector_size__(64)));
-
-/* Unsigned types */
-typedef unsigned char __v64qu __attribute__((__vector_size__(64)));
-typedef unsigned short __v32hu __attribute__((__vector_size__(64)));
-typedef unsigned long long __v8du __attribute__((__vector_size__(64)));
-typedef unsigned int __v16su __attribute__((__vector_size__(64)));
-
-typedef float __m512 __attribute__((__vector_size__(64), __aligned__(64)));
-typedef double __m512d __attribute__((__vector_size__(64), __aligned__(64)));
-typedef long long __m512i __attribute__((__vector_size__(64), __aligned__(64)));
-
-typedef float __m512_u __attribute__((__vector_size__(64), __aligned__(1)));
-typedef double __m512d_u __attribute__((__vector_size__(64), __aligned__(1)));
-typedef long long __m512i_u __attribute__((__vector_size__(64), __aligned__(1)));
-
-typedef unsigned char __mmask8;
-typedef unsigned short __mmask16;
-
-/* Rounding mode macros.  */
-#define _MM_FROUND_TO_NEAREST_INT   0x00
-#define _MM_FROUND_TO_NEG_INF       0x01
-#define _MM_FROUND_TO_POS_INF       0x02
-#define _MM_FROUND_TO_ZERO          0x03
-#define _MM_FROUND_CUR_DIRECTION    0x04
-
-/* Constants for integer comparison predicates */
-typedef enum {
-    _MM_CMPINT_EQ,      /* Equal */
-    _MM_CMPINT_LT,      /* Less than */
-    _MM_CMPINT_LE,      /* Less than or Equal */
-    _MM_CMPINT_UNUSED,
-    _MM_CMPINT_NE,      /* Not Equal */
-    _MM_CMPINT_NLT,     /* Not Less than */
-#define _MM_CMPINT_GE   _MM_CMPINT_NLT  /* Greater than or Equal */
-    _MM_CMPINT_NLE      /* Not Less than or Equal */
-#define _MM_CMPINT_GT   _MM_CMPINT_NLE  /* Greater than */
-} _MM_CMPINT_ENUM;
-
-typedef enum
-{
-  _MM_PERM_AAAA = 0x00, _MM_PERM_AAAB = 0x01, _MM_PERM_AAAC = 0x02,
-  _MM_PERM_AAAD = 0x03, _MM_PERM_AABA = 0x04, _MM_PERM_AABB = 0x05,
-  _MM_PERM_AABC = 0x06, _MM_PERM_AABD = 0x07, _MM_PERM_AACA = 0x08,
-  _MM_PERM_AACB = 0x09, _MM_PERM_AACC = 0x0A, _MM_PERM_AACD = 0x0B,
-  _MM_PERM_AADA = 0x0C, _MM_PERM_AADB = 0x0D, _MM_PERM_AADC = 0x0E,
-  _MM_PERM_AADD = 0x0F, _MM_PERM_ABAA = 0x10, _MM_PERM_ABAB = 0x11,
-  _MM_PERM_ABAC = 0x12, _MM_PERM_ABAD = 0x13, _MM_PERM_ABBA = 0x14,
-  _MM_PERM_ABBB = 0x15, _MM_PERM_ABBC = 0x16, _MM_PERM_ABBD = 0x17,
-  _MM_PERM_ABCA = 0x18, _MM_PERM_ABCB = 0x19, _MM_PERM_ABCC = 0x1A,
-  _MM_PERM_ABCD = 0x1B, _MM_PERM_ABDA = 0x1C, _MM_PERM_ABDB = 0x1D,
-  _MM_PERM_ABDC = 0x1E, _MM_PERM_ABDD = 0x1F, _MM_PERM_ACAA = 0x20,
-  _MM_PERM_ACAB = 0x21, _MM_PERM_ACAC = 0x22, _MM_PERM_ACAD = 0x23,
-  _MM_PERM_ACBA = 0x24, _MM_PERM_ACBB = 0x25, _MM_PERM_ACBC = 0x26,
-  _MM_PERM_ACBD = 0x27, _MM_PERM_ACCA = 0x28, _MM_PERM_ACCB = 0x29,
-  _MM_PERM_ACCC = 0x2A, _MM_PERM_ACCD = 0x2B, _MM_PERM_ACDA = 0x2C,
-  _MM_PERM_ACDB = 0x2D, _MM_PERM_ACDC = 0x2E, _MM_PERM_ACDD = 0x2F,
-  _MM_PERM_ADAA = 0x30, _MM_PERM_ADAB = 0x31, _MM_PERM_ADAC = 0x32,
-  _MM_PERM_ADAD = 0x33, _MM_PERM_ADBA = 0x34, _MM_PERM_ADBB = 0x35,
-  _MM_PERM_ADBC = 0x36, _MM_PERM_ADBD = 0x37, _MM_PERM_ADCA = 0x38,
-  _MM_PERM_ADCB = 0x39, _MM_PERM_ADCC = 0x3A, _MM_PERM_ADCD = 0x3B,
-  _MM_PERM_ADDA = 0x3C, _MM_PERM_ADDB = 0x3D, _MM_PERM_ADDC = 0x3E,
-  _MM_PERM_ADDD = 0x3F, _MM_PERM_BAAA = 0x40, _MM_PERM_BAAB = 0x41,
-  _MM_PERM_BAAC = 0x42, _MM_PERM_BAAD = 0x43, _MM_PERM_BABA = 0x44,
-  _MM_PERM_BABB = 0x45, _MM_PERM_BABC = 0x46, _MM_PERM_BABD = 0x47,
-  _MM_PERM_BACA = 0x48, _MM_PERM_BACB = 0x49, _MM_PERM_BACC = 0x4A,
-  _MM_PERM_BACD = 0x4B, _MM_PERM_BADA = 0x4C, _MM_PERM_BADB = 0x4D,
-  _MM_PERM_BADC = 0x4E, _MM_PERM_BADD = 0x4F, _MM_PERM_BBAA = 0x50,
-  _MM_PERM_BBAB = 0x51, _MM_PERM_BBAC = 0x52, _MM_PERM_BBAD = 0x53,
-  _MM_PERM_BBBA = 0x54, _MM_PERM_BBBB = 0x55, _MM_PERM_BBBC = 0x56,
-  _MM_PERM_BBBD = 0x57, _MM_PERM_BBCA = 0x58, _MM_PERM_BBCB = 0x59,
-  _MM_PERM_BBCC = 0x5A, _MM_PERM_BBCD = 0x5B, _MM_PERM_BBDA = 0x5C,
-  _MM_PERM_BBDB = 0x5D, _MM_PERM_BBDC = 0x5E, _MM_PERM_BBDD = 0x5F,
-  _MM_PERM_BCAA = 0x60, _MM_PERM_BCAB = 0x61, _MM_PERM_BCAC = 0x62,
-  _MM_PERM_BCAD = 0x63, _MM_PERM_BCBA = 0x64, _MM_PERM_BCBB = 0x65,
-  _MM_PERM_BCBC = 0x66, _MM_PERM_BCBD = 0x67, _MM_PERM_BCCA = 0x68,
-  _MM_PERM_BCCB = 0x69, _MM_PERM_BCCC = 0x6A, _MM_PERM_BCCD = 0x6B,
-  _MM_PERM_BCDA = 0x6C, _MM_PERM_BCDB = 0x6D, _MM_PERM_BCDC = 0x6E,
-  _MM_PERM_BCDD = 0x6F, _MM_PERM_BDAA = 0x70, _MM_PERM_BDAB = 0x71,
-  _MM_PERM_BDAC = 0x72, _MM_PERM_BDAD = 0x73, _MM_PERM_BDBA = 0x74,
-  _MM_PERM_BDBB = 0x75, _MM_PERM_BDBC = 0x76, _MM_PERM_BDBD = 0x77,
-  _MM_PERM_BDCA = 0x78, _MM_PERM_BDCB = 0x79, _MM_PERM_BDCC = 0x7A,
-  _MM_PERM_BDCD = 0x7B, _MM_PERM_BDDA = 0x7C, _MM_PERM_BDDB = 0x7D,
-  _MM_PERM_BDDC = 0x7E, _MM_PERM_BDDD = 0x7F, _MM_PERM_CAAA = 0x80,
-  _MM_PERM_CAAB = 0x81, _MM_PERM_CAAC = 0x82, _MM_PERM_CAAD = 0x83,
-  _MM_PERM_CABA = 0x84, _MM_PERM_CABB = 0x85, _MM_PERM_CABC = 0x86,
-  _MM_PERM_CABD = 0x87, _MM_PERM_CACA = 0x88, _MM_PERM_CACB = 0x89,
-  _MM_PERM_CACC = 0x8A, _MM_PERM_CACD = 0x8B, _MM_PERM_CADA = 0x8C,
-  _MM_PERM_CADB = 0x8D, _MM_PERM_CADC = 0x8E, _MM_PERM_CADD = 0x8F,
-  _MM_PERM_CBAA = 0x90, _MM_PERM_CBAB = 0x91, _MM_PERM_CBAC = 0x92,
-  _MM_PERM_CBAD = 0x93, _MM_PERM_CBBA = 0x94, _MM_PERM_CBBB = 0x95,
-  _MM_PERM_CBBC = 0x96, _MM_PERM_CBBD = 0x97, _MM_PERM_CBCA = 0x98,
-  _MM_PERM_CBCB = 0x99, _MM_PERM_CBCC = 0x9A, _MM_PERM_CBCD = 0x9B,
-  _MM_PERM_CBDA = 0x9C, _MM_PERM_CBDB = 0x9D, _MM_PERM_CBDC = 0x9E,
-  _MM_PERM_CBDD = 0x9F, _MM_PERM_CCAA = 0xA0, _MM_PERM_CCAB = 0xA1,
-  _MM_PERM_CCAC = 0xA2, _MM_PERM_CCAD = 0xA3, _MM_PERM_CCBA = 0xA4,
-  _MM_PERM_CCBB = 0xA5, _MM_PERM_CCBC = 0xA6, _MM_PERM_CCBD = 0xA7,
-  _MM_PERM_CCCA = 0xA8, _MM_PERM_CCCB = 0xA9, _MM_PERM_CCCC = 0xAA,
-  _MM_PERM_CCCD = 0xAB, _MM_PERM_CCDA = 0xAC, _MM_PERM_CCDB = 0xAD,
-  _MM_PERM_CCDC = 0xAE, _MM_PERM_CCDD = 0xAF, _MM_PERM_CDAA = 0xB0,
-  _MM_PERM_CDAB = 0xB1, _MM_PERM_CDAC = 0xB2, _MM_PERM_CDAD = 0xB3,
-  _MM_PERM_CDBA = 0xB4, _MM_PERM_CDBB = 0xB5, _MM_PERM_CDBC = 0xB6,
-  _MM_PERM_CDBD = 0xB7, _MM_PERM_CDCA = 0xB8, _MM_PERM_CDCB = 0xB9,
-  _MM_PERM_CDCC = 0xBA, _MM_PERM_CDCD = 0xBB, _MM_PERM_CDDA = 0xBC,
-  _MM_PERM_CDDB = 0xBD, _MM_PERM_CDDC = 0xBE, _MM_PERM_CDDD = 0xBF,
-  _MM_PERM_DAAA = 0xC0, _MM_PERM_DAAB = 0xC1, _MM_PERM_DAAC = 0xC2,
-  _MM_PERM_DAAD = 0xC3, _MM_PERM_DABA = 0xC4, _MM_PERM_DABB = 0xC5,
-  _MM_PERM_DABC = 0xC6, _MM_PERM_DABD = 0xC7, _MM_PERM_DACA = 0xC8,
-  _MM_PERM_DACB = 0xC9, _MM_PERM_DACC = 0xCA, _MM_PERM_DACD = 0xCB,
-  _MM_PERM_DADA = 0xCC, _MM_PERM_DADB = 0xCD, _MM_PERM_DADC = 0xCE,
-  _MM_PERM_DADD = 0xCF, _MM_PERM_DBAA = 0xD0, _MM_PERM_DBAB = 0xD1,
-  _MM_PERM_DBAC = 0xD2, _MM_PERM_DBAD = 0xD3, _MM_PERM_DBBA = 0xD4,
-  _MM_PERM_DBBB = 0xD5, _MM_PERM_DBBC = 0xD6, _MM_PERM_DBBD = 0xD7,
-  _MM_PERM_DBCA = 0xD8, _MM_PERM_DBCB = 0xD9, _MM_PERM_DBCC = 0xDA,
-  _MM_PERM_DBCD = 0xDB, _MM_PERM_DBDA = 0xDC, _MM_PERM_DBDB = 0xDD,
-  _MM_PERM_DBDC = 0xDE, _MM_PERM_DBDD = 0xDF, _MM_PERM_DCAA = 0xE0,
-  _MM_PERM_DCAB = 0xE1, _MM_PERM_DCAC = 0xE2, _MM_PERM_DCAD = 0xE3,
-  _MM_PERM_DCBA = 0xE4, _MM_PERM_DCBB = 0xE5, _MM_PERM_DCBC = 0xE6,
-  _MM_PERM_DCBD = 0xE7, _MM_PERM_DCCA = 0xE8, _MM_PERM_DCCB = 0xE9,
-  _MM_PERM_DCCC = 0xEA, _MM_PERM_DCCD = 0xEB, _MM_PERM_DCDA = 0xEC,
-  _MM_PERM_DCDB = 0xED, _MM_PERM_DCDC = 0xEE, _MM_PERM_DCDD = 0xEF,
-  _MM_PERM_DDAA = 0xF0, _MM_PERM_DDAB = 0xF1, _MM_PERM_DDAC = 0xF2,
-  _MM_PERM_DDAD = 0xF3, _MM_PERM_DDBA = 0xF4, _MM_PERM_DDBB = 0xF5,
-  _MM_PERM_DDBC = 0xF6, _MM_PERM_DDBD = 0xF7, _MM_PERM_DDCA = 0xF8,
-  _MM_PERM_DDCB = 0xF9, _MM_PERM_DDCC = 0xFA, _MM_PERM_DDCD = 0xFB,
-  _MM_PERM_DDDA = 0xFC, _MM_PERM_DDDB = 0xFD, _MM_PERM_DDDC = 0xFE,
-  _MM_PERM_DDDD = 0xFF
-} _MM_PERM_ENUM;
-
-typedef enum
-{
-  _MM_MANT_NORM_1_2,    /* interval [1, 2)      */
-  _MM_MANT_NORM_p5_2,   /* interval [0.5, 2)    */
-  _MM_MANT_NORM_p5_1,   /* interval [0.5, 1)    */
-  _MM_MANT_NORM_p75_1p5   /* interval [0.75, 1.5) */
-} _MM_MANTISSA_NORM_ENUM;
-
-typedef enum
-{
-  _MM_MANT_SIGN_src,    /* sign = sign(SRC)     */
-  _MM_MANT_SIGN_zero,   /* sign = 0             */
-  _MM_MANT_SIGN_nan   /* DEST = NaN if sign(SRC) = 1 */
-} _MM_MANTISSA_SIGN_ENUM;
-
-/* Define the default attributes for the functions in this file. */
-#define __DEFAULT_FN_ATTRS512 __attribute__((__always_inline__, __nodebug__, __target__("avx512f"), __min_vector_width__(512)))
-#define __DEFAULT_FN_ATTRS128 __attribute__((__always_inline__, __nodebug__, __target__("avx512f"), __min_vector_width__(128)))
-#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("avx512f")))
-
-/* Create vectors with repeated elements */
-
-static  __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_setzero_si512(void)
-{
-  return __extension__ (__m512i)(__v8di){ 0, 0, 0, 0, 0, 0, 0, 0 };
-}
-
-#define _mm512_setzero_epi32 _mm512_setzero_si512
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_undefined_pd(void)
-{
-  return (__m512d)__builtin_ia32_undef512();
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_undefined(void)
-{
-  return (__m512)__builtin_ia32_undef512();
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_undefined_ps(void)
-{
-  return (__m512)__builtin_ia32_undef512();
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_undefined_epi32(void)
-{
-  return (__m512i)__builtin_ia32_undef512();
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_broadcastd_epi32 (__m128i __A)
-{
-  return (__m512i)__builtin_shufflevector((__v4si) __A, (__v4si) __A,
-                                          0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcastd_epi32 (__m512i __O, __mmask16 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__M,
-                                             (__v16si) _mm512_broadcastd_epi32(__A),
-                                             (__v16si) __O);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcastd_epi32 (__mmask16 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__M,
-                                             (__v16si) _mm512_broadcastd_epi32(__A),
-                                             (__v16si) _mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_broadcastq_epi64 (__m128i __A)
-{
-  return (__m512i)__builtin_shufflevector((__v2di) __A, (__v2di) __A,
-                                          0, 0, 0, 0, 0, 0, 0, 0);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcastq_epi64 (__m512i __O, __mmask8 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__M,
-                                             (__v8di) _mm512_broadcastq_epi64(__A),
-                                             (__v8di) __O);
-
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcastq_epi64 (__mmask8 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__M,
-                                             (__v8di) _mm512_broadcastq_epi64(__A),
-                                             (__v8di) _mm512_setzero_si512());
-}
-
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_setzero_ps(void)
-{
-  return __extension__ (__m512){ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
-                                 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 };
-}
-
-#define _mm512_setzero _mm512_setzero_ps
-
-static  __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_setzero_pd(void)
-{
-  return __extension__ (__m512d){ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 };
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_set1_ps(float __w)
-{
-  return __extension__ (__m512){ __w, __w, __w, __w, __w, __w, __w, __w,
-                                 __w, __w, __w, __w, __w, __w, __w, __w  };
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_set1_pd(double __w)
-{
-  return __extension__ (__m512d){ __w, __w, __w, __w, __w, __w, __w, __w };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set1_epi8(char __w)
-{
-  return __extension__ (__m512i)(__v64qi){
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w  };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set1_epi16(short __w)
-{
-  return __extension__ (__m512i)(__v32hi){
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w,
-    __w, __w, __w, __w, __w, __w, __w, __w };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set1_epi32(int __s)
-{
-  return __extension__ (__m512i)(__v16si){
-    __s, __s, __s, __s, __s, __s, __s, __s,
-    __s, __s, __s, __s, __s, __s, __s, __s };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_set1_epi32(__mmask16 __M, int __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__M,
-                                             (__v16si)_mm512_set1_epi32(__A),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set1_epi64(long long __d)
-{
-  return __extension__(__m512i)(__v8di){ __d, __d, __d, __d, __d, __d, __d, __d };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_set1_epi64(__mmask8 __M, long long __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__M,
-                                             (__v8di)_mm512_set1_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_broadcastss_ps(__m128 __A)
-{
-  return (__m512)__builtin_shufflevector((__v4sf) __A, (__v4sf) __A,
-                                         0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set4_epi32 (int __A, int __B, int __C, int __D)
-{
-  return __extension__ (__m512i)(__v16si)
-   { __D, __C, __B, __A, __D, __C, __B, __A,
-     __D, __C, __B, __A, __D, __C, __B, __A };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set4_epi64 (long long __A, long long __B, long long __C,
-       long long __D)
-{
-  return __extension__ (__m512i) (__v8di)
-   { __D, __C, __B, __A, __D, __C, __B, __A };
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_set4_pd (double __A, double __B, double __C, double __D)
-{
-  return __extension__ (__m512d)
-   { __D, __C, __B, __A, __D, __C, __B, __A };
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_set4_ps (float __A, float __B, float __C, float __D)
-{
-  return __extension__ (__m512)
-   { __D, __C, __B, __A, __D, __C, __B, __A,
-     __D, __C, __B, __A, __D, __C, __B, __A };
-}
-
-#define _mm512_setr4_epi32(e0,e1,e2,e3)               \
-  _mm512_set4_epi32((e3),(e2),(e1),(e0))
-
-#define _mm512_setr4_epi64(e0,e1,e2,e3)               \
-  _mm512_set4_epi64((e3),(e2),(e1),(e0))
-
-#define _mm512_setr4_pd(e0,e1,e2,e3)                \
-  _mm512_set4_pd((e3),(e2),(e1),(e0))
-
-#define _mm512_setr4_ps(e0,e1,e2,e3)                \
-  _mm512_set4_ps((e3),(e2),(e1),(e0))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_broadcastsd_pd(__m128d __A)
-{
-  return (__m512d)__builtin_shufflevector((__v2df) __A, (__v2df) __A,
-                                          0, 0, 0, 0, 0, 0, 0, 0);
-}
-
-/* Cast between vector types */
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_castpd256_pd512(__m256d __a)
-{
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, -1, -1, -1, -1);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_castps256_ps512(__m256 __a)
-{
-  return __builtin_shufflevector(__a, __a, 0,  1,  2,  3,  4,  5,  6,  7,
-                                          -1, -1, -1, -1, -1, -1, -1, -1);
-}
-
-static __inline __m128d __DEFAULT_FN_ATTRS512
-_mm512_castpd512_pd128(__m512d __a)
-{
-  return __builtin_shufflevector(__a, __a, 0, 1);
-}
-
-static __inline __m256d __DEFAULT_FN_ATTRS512
-_mm512_castpd512_pd256 (__m512d __A)
-{
-  return __builtin_shufflevector(__A, __A, 0, 1, 2, 3);
-}
-
-static __inline __m128 __DEFAULT_FN_ATTRS512
-_mm512_castps512_ps128(__m512 __a)
-{
-  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3);
-}
-
-static __inline __m256 __DEFAULT_FN_ATTRS512
-_mm512_castps512_ps256 (__m512 __A)
-{
-  return __builtin_shufflevector(__A, __A, 0, 1, 2, 3, 4, 5, 6, 7);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_castpd_ps (__m512d __A)
-{
-  return (__m512) (__A);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_castpd_si512 (__m512d __A)
-{
-  return (__m512i) (__A);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_castpd128_pd512 (__m128d __A)
-{
-  return __builtin_shufflevector( __A, __A, 0, 1, -1, -1, -1, -1, -1, -1);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_castps_pd (__m512 __A)
-{
-  return (__m512d) (__A);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_castps_si512 (__m512 __A)
-{
-  return (__m512i) (__A);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_castps128_ps512 (__m128 __A)
-{
-    return  __builtin_shufflevector( __A, __A, 0, 1, 2, 3, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_castsi128_si512 (__m128i __A)
-{
-   return  __builtin_shufflevector( __A, __A, 0, 1, -1, -1, -1, -1, -1, -1);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_castsi256_si512 (__m256i __A)
-{
-   return  __builtin_shufflevector( __A, __A, 0, 1, 2, 3, -1, -1, -1, -1);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_castsi512_ps (__m512i __A)
-{
-  return (__m512) (__A);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_castsi512_pd (__m512i __A)
-{
-  return (__m512d) (__A);
-}
-
-static __inline __m128i __DEFAULT_FN_ATTRS512
-_mm512_castsi512_si128 (__m512i __A)
-{
-  return (__m128i)__builtin_shufflevector(__A, __A , 0, 1);
-}
-
-static __inline __m256i __DEFAULT_FN_ATTRS512
-_mm512_castsi512_si256 (__m512i __A)
-{
-  return (__m256i)__builtin_shufflevector(__A, __A , 0, 1, 2, 3);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_int2mask(int __a)
-{
-  return (__mmask16)__a;
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm512_mask2int(__mmask16 __a)
-{
-  return (int)__a;
-}
-
-/// Constructs a 512-bit floating-point vector of [8 x double] from a
-///    128-bit floating-point vector of [2 x double]. The lower 128 bits
-///    contain the value of the source vector. The upper 384 bits are set
-///    to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 512-bit floating-point vector of [8 x double]. The lower 128 bits
-///    contain the value of the parameter. The upper 384 bits are set to zero.
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_zextpd128_pd512(__m128d __a)
-{
-  return __builtin_shufflevector((__v2df)__a, (__v2df)_mm_setzero_pd(), 0, 1, 2, 3, 2, 3, 2, 3);
-}
-
-/// Constructs a 512-bit floating-point vector of [8 x double] from a
-///    256-bit floating-point vector of [4 x double]. The lower 256 bits
-///    contain the value of the source vector. The upper 256 bits are set
-///    to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 256-bit vector of [4 x double].
-/// \returns A 512-bit floating-point vector of [8 x double]. The lower 256 bits
-///    contain the value of the parameter. The upper 256 bits are set to zero.
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_zextpd256_pd512(__m256d __a)
-{
-  return __builtin_shufflevector((__v4df)__a, (__v4df)_mm256_setzero_pd(), 0, 1, 2, 3, 4, 5, 6, 7);
-}
-
-/// Constructs a 512-bit floating-point vector of [16 x float] from a
-///    128-bit floating-point vector of [4 x float]. The lower 128 bits contain
-///    the value of the source vector. The upper 384 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 512-bit floating-point vector of [16 x float]. The lower 128 bits
-///    contain the value of the parameter. The upper 384 bits are set to zero.
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_zextps128_ps512(__m128 __a)
-{
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)_mm_setzero_ps(), 0, 1, 2, 3, 4, 5, 6, 7, 4, 5, 6, 7, 4, 5, 6, 7);
-}
-
-/// Constructs a 512-bit floating-point vector of [16 x float] from a
-///    256-bit floating-point vector of [8 x float]. The lower 256 bits contain
-///    the value of the source vector. The upper 256 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 256-bit vector of [8 x float].
-/// \returns A 512-bit floating-point vector of [16 x float]. The lower 256 bits
-///    contain the value of the parameter. The upper 256 bits are set to zero.
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_zextps256_ps512(__m256 __a)
-{
-  return __builtin_shufflevector((__v8sf)__a, (__v8sf)_mm256_setzero_ps(), 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-}
-
-/// Constructs a 512-bit integer vector from a 128-bit integer vector.
-///    The lower 128 bits contain the value of the source vector. The upper
-///    384 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \returns A 512-bit integer vector. The lower 128 bits contain the value of
-///    the parameter. The upper 384 bits are set to zero.
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_zextsi128_si512(__m128i __a)
-{
-  return __builtin_shufflevector((__v2di)__a, (__v2di)_mm_setzero_si128(), 0, 1, 2, 3, 2, 3, 2, 3);
-}
-
-/// Constructs a 512-bit integer vector from a 256-bit integer vector.
-///    The lower 256 bits contain the value of the source vector. The upper
-///    256 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 256-bit integer vector.
-/// \returns A 512-bit integer vector. The lower 256 bits contain the value of
-///    the parameter. The upper 256 bits are set to zero.
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_zextsi256_si512(__m256i __a)
-{
-  return __builtin_shufflevector((__v4di)__a, (__v4di)_mm256_setzero_si256(), 0, 1, 2, 3, 4, 5, 6, 7);
-}
-
-/* Bitwise operators */
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_and_epi32(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v16su)__a & (__v16su)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_and_epi32(__m512i __src, __mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__k,
-                (__v16si) _mm512_and_epi32(__a, __b),
-                (__v16si) __src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_and_epi32(__mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i) _mm512_mask_and_epi32(_mm512_setzero_si512 (),
-                                         __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_and_epi64(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a & (__v8du)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_and_epi64(__m512i __src, __mmask8 __k, __m512i __a, __m512i __b)
-{
-    return (__m512i) __builtin_ia32_selectq_512 ((__mmask8) __k,
-                (__v8di) _mm512_and_epi64(__a, __b),
-                (__v8di) __src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_and_epi64(__mmask8 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i) _mm512_mask_and_epi64(_mm512_setzero_si512 (),
-                                         __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_andnot_si512 (__m512i __A, __m512i __B)
-{
-  return (__m512i)(~(__v8du)__A & (__v8du)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_andnot_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i)(~(__v16su)__A & (__v16su)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_andnot_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_andnot_epi32(__A, __B),
-                                         (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_andnot_epi32(__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)_mm512_mask_andnot_epi32(_mm512_setzero_si512(),
-                                           __U, __A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_andnot_epi64(__m512i __A, __m512i __B)
-{
-  return (__m512i)(~(__v8du)__A & (__v8du)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_andnot_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_andnot_epi64(__A, __B),
-                                          (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_andnot_epi64(__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)_mm512_mask_andnot_epi64(_mm512_setzero_si512(),
-                                           __U, __A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_or_epi32(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v16su)__a | (__v16su)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_or_epi32(__m512i __src, __mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__k,
-                                             (__v16si)_mm512_or_epi32(__a, __b),
-                                             (__v16si)__src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_or_epi32(__mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)_mm512_mask_or_epi32(_mm512_setzero_si512(), __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_or_epi64(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a | (__v8du)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_or_epi64(__m512i __src, __mmask8 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__k,
-                                             (__v8di)_mm512_or_epi64(__a, __b),
-                                             (__v8di)__src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_or_epi64(__mmask8 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)_mm512_mask_or_epi64(_mm512_setzero_si512(), __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_xor_epi32(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v16su)__a ^ (__v16su)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_xor_epi32(__m512i __src, __mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__k,
-                                            (__v16si)_mm512_xor_epi32(__a, __b),
-                                            (__v16si)__src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_xor_epi32(__mmask16 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)_mm512_mask_xor_epi32(_mm512_setzero_si512(), __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_xor_epi64(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a ^ (__v8du)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_xor_epi64(__m512i __src, __mmask8 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__k,
-                                             (__v8di)_mm512_xor_epi64(__a, __b),
-                                             (__v8di)__src);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_xor_epi64(__mmask8 __k, __m512i __a, __m512i __b)
-{
-  return (__m512i)_mm512_mask_xor_epi64(_mm512_setzero_si512(), __k, __a, __b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_and_si512(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a & (__v8du)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_or_si512(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a | (__v8du)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_xor_si512(__m512i __a, __m512i __b)
-{
-  return (__m512i)((__v8du)__a ^ (__v8du)__b);
-}
-
-/* Arithmetic */
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_add_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)((__v8df)__a + (__v8df)__b);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_add_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)((__v16sf)__a + (__v16sf)__b);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_mul_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)((__v8df)__a * (__v8df)__b);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_mul_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)((__v16sf)__a * (__v16sf)__b);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_sub_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)((__v8df)__a - (__v8df)__b);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_sub_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)((__v16sf)__a - (__v16sf)__b);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_add_epi64 (__m512i __A, __m512i __B)
-{
-  return (__m512i) ((__v8du) __A + (__v8du) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_add_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_add_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_add_epi64(__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_add_epi64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sub_epi64 (__m512i __A, __m512i __B)
-{
-  return (__m512i) ((__v8du) __A - (__v8du) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sub_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_sub_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sub_epi64(__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_sub_epi64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_add_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i) ((__v16su) __A + (__v16su) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_add_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_add_epi32(__A, __B),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_add_epi32 (__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_add_epi32(__A, __B),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sub_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i) ((__v16su) __A - (__v16su) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sub_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_sub_epi32(__A, __B),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sub_epi32(__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_sub_epi32(__A, __B),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-#define _mm512_max_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_maxpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_max_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_max_round_pd((A), (B), (R)), \
-                                   (__v8df)(W))
-
-#define _mm512_maskz_max_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_max_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-static  __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_max_pd(__m512d __A, __m512d __B)
-{
-  return (__m512d) __builtin_ia32_maxpd512((__v8df) __A, (__v8df) __B,
-                                           _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_max_pd (__m512d __W, __mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_max_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_pd (__mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_max_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-#define _mm512_max_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_maxps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_max_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_max_round_ps((A), (B), (R)), \
-                                  (__v16sf)(W))
-
-#define _mm512_maskz_max_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_max_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_max_ps(__m512 __A, __m512 __B)
-{
-  return (__m512) __builtin_ia32_maxps512((__v16sf) __A, (__v16sf) __B,
-                                          _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_max_ps (__m512 __W, __mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_max_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_ps (__mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_max_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_max_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  return (__m128) __builtin_ia32_maxss_round_mask ((__v4sf) __A,
-                (__v4sf) __B,
-                (__v4sf) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_max_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  return (__m128) __builtin_ia32_maxss_round_mask ((__v4sf) __A,
-                (__v4sf) __B,
-                (__v4sf)  _mm_setzero_ps (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_max_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_maxss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_max_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_maxss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_max_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_maxss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_max_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  return (__m128d) __builtin_ia32_maxsd_round_mask ((__v2df) __A,
-                (__v2df) __B,
-                (__v2df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_max_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  return (__m128d) __builtin_ia32_maxsd_round_mask ((__v2df) __A,
-                (__v2df) __B,
-                (__v2df)  _mm_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_max_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_maxsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_max_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_maxsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_max_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_maxsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline __m512i
-__DEFAULT_FN_ATTRS512
-_mm512_max_epi32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pmaxsd512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_max_epi32 (__m512i __W, __mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_max_epi32(__A, __B),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_epi32 (__mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_max_epi32(__A, __B),
-                                            (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_max_epu32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pmaxud512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_max_epu32 (__m512i __W, __mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_max_epu32(__A, __B),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_epu32 (__mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_max_epu32(__A, __B),
-                                            (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_max_epi64(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pmaxsq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_max_epi64 (__m512i __W, __mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_max_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_epi64 (__mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_max_epi64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_max_epu64(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pmaxuq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_max_epu64 (__m512i __W, __mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_max_epu64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_max_epu64 (__mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_max_epu64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-#define _mm512_min_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_minpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_min_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_min_round_pd((A), (B), (R)), \
-                                   (__v8df)(W))
-
-#define _mm512_maskz_min_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_min_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-static  __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_min_pd(__m512d __A, __m512d __B)
-{
-  return (__m512d) __builtin_ia32_minpd512((__v8df) __A, (__v8df) __B,
-                                           _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_min_pd (__m512d __W, __mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_min_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_pd (__mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_min_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-#define _mm512_min_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_minps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_min_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_min_round_ps((A), (B), (R)), \
-                                  (__v16sf)(W))
-
-#define _mm512_maskz_min_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_min_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_min_ps(__m512 __A, __m512 __B)
-{
-  return (__m512) __builtin_ia32_minps512((__v16sf) __A, (__v16sf) __B,
-                                          _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_min_ps (__m512 __W, __mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_min_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_ps (__mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_min_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_min_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  return (__m128) __builtin_ia32_minss_round_mask ((__v4sf) __A,
-                (__v4sf) __B,
-                (__v4sf) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_min_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  return (__m128) __builtin_ia32_minss_round_mask ((__v4sf) __A,
-                (__v4sf) __B,
-                (__v4sf)  _mm_setzero_ps (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_min_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_minss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_min_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_minss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_min_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_minss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_min_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  return (__m128d) __builtin_ia32_minsd_round_mask ((__v2df) __A,
-                (__v2df) __B,
-                (__v2df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_min_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  return (__m128d) __builtin_ia32_minsd_round_mask ((__v2df) __A,
-                (__v2df) __B,
-                (__v2df)  _mm_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_min_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_minsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_min_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_minsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_min_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_minsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline __m512i
-__DEFAULT_FN_ATTRS512
-_mm512_min_epi32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pminsd512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_min_epi32 (__m512i __W, __mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_min_epi32(__A, __B),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_epi32 (__mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_min_epi32(__A, __B),
-                                            (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_min_epu32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pminud512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_min_epu32 (__m512i __W, __mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_min_epu32(__A, __B),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_epu32 (__mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                            (__v16si)_mm512_min_epu32(__A, __B),
-                                            (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_min_epi64(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pminsq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_min_epi64 (__m512i __W, __mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_min_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_epi64 (__mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_min_epi64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_min_epu64(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_pminuq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_min_epu64 (__m512i __W, __mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_min_epu64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_min_epu64 (__mmask8 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_min_epu64(__A, __B),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mul_epi32(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_pmuldq512((__v16si)__X, (__v16si) __Y);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mul_epi32(__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_mul_epi32(__X, __Y),
-                                             (__v8di)__W);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_mul_epi32(__mmask8 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_mul_epi32(__X, __Y),
-                                             (__v8di)_mm512_setzero_si512 ());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mul_epu32(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_pmuludq512((__v16si)__X, (__v16si)__Y);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mul_epu32(__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_mul_epu32(__X, __Y),
-                                             (__v8di)__W);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_mul_epu32(__mmask8 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                             (__v8di)_mm512_mul_epu32(__X, __Y),
-                                             (__v8di)_mm512_setzero_si512 ());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mullo_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i) ((__v16su) __A * (__v16su) __B);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_mullo_epi32(__mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                             (__v16si)_mm512_mullo_epi32(__A, __B),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mullo_epi32(__m512i __W, __mmask16 __M, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                             (__v16si)_mm512_mullo_epi32(__A, __B),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mullox_epi64 (__m512i __A, __m512i __B) {
-  return (__m512i) ((__v8du) __A * (__v8du) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mullox_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m512i __B) {
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_mullox_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-#define _mm512_sqrt_round_pd(A, R) \
-  (__m512d)__builtin_ia32_sqrtpd512((__v8df)(__m512d)(A), (int)(R))
-
-#define _mm512_mask_sqrt_round_pd(W, U, A, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_sqrt_round_pd((A), (R)), \
-                                       (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_sqrt_round_pd(U, A, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_sqrt_round_pd((A), (R)), \
-                                       (__v8df)_mm512_setzero_pd())
-
-static  __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_sqrt_pd(__m512d __A)
-{
-  return (__m512d)__builtin_ia32_sqrtpd512((__v8df)__A,
-                                           _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_sqrt_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_sqrt_pd(__A),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_sqrt_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                              (__v8df)_mm512_sqrt_pd(__A),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-#define _mm512_sqrt_round_ps(A, R) \
-  (__m512)__builtin_ia32_sqrtps512((__v16sf)(__m512)(A), (int)(R))
-
-#define _mm512_mask_sqrt_round_ps(W, U, A, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_sqrt_round_ps((A), (R)), \
-                                      (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_sqrt_round_ps(U, A, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_sqrt_round_ps((A), (R)), \
-                                      (__v16sf)_mm512_setzero_ps())
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_sqrt_ps(__m512 __A)
-{
-  return (__m512)__builtin_ia32_sqrtps512((__v16sf)__A,
-                                          _MM_FROUND_CUR_DIRECTION);
-}
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_sqrt_ps(__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_sqrt_ps(__A),
-                                             (__v16sf)__W);
-}
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_sqrt_ps( __mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                             (__v16sf)_mm512_sqrt_ps(__A),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static  __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_rsqrt14_pd(__m512d __A)
-{
-  return (__m512d) __builtin_ia32_rsqrt14pd512_mask ((__v8df) __A,
-                 (__v8df)
-                 _mm512_setzero_pd (),
-                 (__mmask8) -1);}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_rsqrt14_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rsqrt14pd512_mask ((__v8df) __A,
-                  (__v8df) __W,
-                  (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_rsqrt14_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rsqrt14pd512_mask ((__v8df) __A,
-                  (__v8df)
-                  _mm512_setzero_pd (),
-                  (__mmask8) __U);
-}
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_rsqrt14_ps(__m512 __A)
-{
-  return (__m512) __builtin_ia32_rsqrt14ps512_mask ((__v16sf) __A,
-                (__v16sf)
-                _mm512_setzero_ps (),
-                (__mmask16) -1);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_rsqrt14_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rsqrt14ps512_mask ((__v16sf) __A,
-                 (__v16sf) __W,
-                 (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_rsqrt14_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rsqrt14ps512_mask ((__v16sf) __A,
-                 (__v16sf)
-                 _mm512_setzero_ps (),
-                 (__mmask16) __U);
-}
-
-static  __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_rsqrt14_ss(__m128 __A, __m128 __B)
-{
-  return (__m128) __builtin_ia32_rsqrt14ss_mask ((__v4sf) __A,
-             (__v4sf) __B,
-             (__v4sf)
-             _mm_setzero_ps (),
-             (__mmask8) -1);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_rsqrt14_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_rsqrt14ss_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) __W,
-          (__mmask8) __U);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_rsqrt14_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_rsqrt14ss_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) _mm_setzero_ps (),
-          (__mmask8) __U);
-}
-
-static  __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_rsqrt14_sd(__m128d __A, __m128d __B)
-{
-  return (__m128d) __builtin_ia32_rsqrt14sd_mask ((__v2df) __A,
-              (__v2df) __B,
-              (__v2df)
-              _mm_setzero_pd (),
-              (__mmask8) -1);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_rsqrt14_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_rsqrt14sd_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) __W,
-          (__mmask8) __U);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_rsqrt14_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_rsqrt14sd_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) _mm_setzero_pd (),
-          (__mmask8) __U);
-}
-
-static  __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_rcp14_pd(__m512d __A)
-{
-  return (__m512d) __builtin_ia32_rcp14pd512_mask ((__v8df) __A,
-               (__v8df)
-               _mm512_setzero_pd (),
-               (__mmask8) -1);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_rcp14_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rcp14pd512_mask ((__v8df) __A,
-                (__v8df) __W,
-                (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_rcp14_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rcp14pd512_mask ((__v8df) __A,
-                (__v8df)
-                _mm512_setzero_pd (),
-                (__mmask8) __U);
-}
-
-static  __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_rcp14_ps(__m512 __A)
-{
-  return (__m512) __builtin_ia32_rcp14ps512_mask ((__v16sf) __A,
-              (__v16sf)
-              _mm512_setzero_ps (),
-              (__mmask16) -1);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_rcp14_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rcp14ps512_mask ((__v16sf) __A,
-                   (__v16sf) __W,
-                   (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_rcp14_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rcp14ps512_mask ((__v16sf) __A,
-                   (__v16sf)
-                   _mm512_setzero_ps (),
-                   (__mmask16) __U);
-}
-
-static  __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_rcp14_ss(__m128 __A, __m128 __B)
-{
-  return (__m128) __builtin_ia32_rcp14ss_mask ((__v4sf) __A,
-                 (__v4sf) __B,
-                 (__v4sf)
-                 _mm_setzero_ps (),
-                 (__mmask8) -1);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_rcp14_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_rcp14ss_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) __W,
-          (__mmask8) __U);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_rcp14_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_rcp14ss_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) _mm_setzero_ps (),
-          (__mmask8) __U);
-}
-
-static  __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_rcp14_sd(__m128d __A, __m128d __B)
-{
-  return (__m128d) __builtin_ia32_rcp14sd_mask ((__v2df) __A,
-            (__v2df) __B,
-            (__v2df)
-            _mm_setzero_pd (),
-            (__mmask8) -1);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_rcp14_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_rcp14sd_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) __W,
-          (__mmask8) __U);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_rcp14_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_rcp14sd_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) _mm_setzero_pd (),
-          (__mmask8) __U);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_floor_ps(__m512 __A)
-{
-  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
-                                                  _MM_FROUND_FLOOR,
-                                                  (__v16sf) __A, -1,
-                                                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_floor_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
-                   _MM_FROUND_FLOOR,
-                   (__v16sf) __W, __U,
-                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_floor_pd(__m512d __A)
-{
-  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
-                                                   _MM_FROUND_FLOOR,
-                                                   (__v8df) __A, -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_floor_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
-                _MM_FROUND_FLOOR,
-                (__v8df) __W, __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_ceil_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
-                   _MM_FROUND_CEIL,
-                   (__v16sf) __W, __U,
-                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_ceil_ps(__m512 __A)
-{
-  return (__m512) __builtin_ia32_rndscaleps_mask ((__v16sf) __A,
-                                                  _MM_FROUND_CEIL,
-                                                  (__v16sf) __A, -1,
-                                                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_ceil_pd(__m512d __A)
-{
-  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
-                                                   _MM_FROUND_CEIL,
-                                                   (__v8df) __A, -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_ceil_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_rndscalepd_mask ((__v8df) __A,
-                _MM_FROUND_CEIL,
-                (__v8df) __W, __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_abs_epi64(__m512i __A)
-{
-  return (__m512i)__builtin_ia32_pabsq512((__v8di)__A);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_abs_epi64 (__m512i __W, __mmask8 __U, __m512i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_abs_epi64(__A),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_abs_epi64 (__mmask8 __U, __m512i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_abs_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_abs_epi32(__m512i __A)
-{
-  return (__m512i)__builtin_ia32_pabsd512((__v16si) __A);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_abs_epi32 (__m512i __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                             (__v16si)_mm512_abs_epi32(__A),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_abs_epi32 (__mmask16 __U, __m512i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                             (__v16si)_mm512_abs_epi32(__A),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_add_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_add_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, __W);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_add_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_add_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, _mm_setzero_ps());
-}
-
-#define _mm_add_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_addss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_add_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_addss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_add_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_addss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_add_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_add_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, __W);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_add_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_add_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, _mm_setzero_pd());
-}
-#define _mm_add_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_addsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_add_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_addsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_add_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_addsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_add_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_add_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_add_pd(__mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_add_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_add_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_add_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_add_ps(__mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_add_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-#define _mm512_add_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_addpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_add_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_add_round_pd((A), (B), (R)), \
-                                   (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_add_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_add_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-#define _mm512_add_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_addps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_add_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_add_round_ps((A), (B), (R)), \
-                                  (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_add_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_add_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_sub_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_sub_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, __W);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_sub_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_sub_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, _mm_setzero_ps());
-}
-#define _mm_sub_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_subss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_sub_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_subss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_sub_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_subss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_sub_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_sub_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, __W);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_sub_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_sub_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, _mm_setzero_pd());
-}
-
-#define _mm_sub_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_subsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_sub_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_subsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_sub_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_subsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_sub_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_sub_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_sub_pd(__mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_sub_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_sub_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_sub_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_sub_ps(__mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_sub_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-#define _mm512_sub_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_subpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_sub_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_sub_round_pd((A), (B), (R)), \
-                                   (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_sub_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_sub_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-#define _mm512_sub_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_subps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_sub_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_sub_round_ps((A), (B), (R)), \
-                                  (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_sub_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_sub_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_mul_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_mul_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, __W);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_mul_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_mul_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, _mm_setzero_ps());
-}
-#define _mm_mul_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_mulss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_mul_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_mulss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_mul_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_mulss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_mul_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_mul_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, __W);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_mul_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_mul_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, _mm_setzero_pd());
-}
-
-#define _mm_mul_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_mulsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_mul_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_mulsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_mul_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_mulsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_mul_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_mul_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_mul_pd(__mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_mul_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_mul_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_mul_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_mul_ps(__mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_mul_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-#define _mm512_mul_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_mulpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_mul_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_mul_round_pd((A), (B), (R)), \
-                                   (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_mul_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_mul_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-#define _mm512_mul_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_mulps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_mul_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_mul_round_ps((A), (B), (R)), \
-                                  (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_mul_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_mul_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_div_ss(__m128 __W, __mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_div_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, __W);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_div_ss(__mmask8 __U,__m128 __A, __m128 __B) {
-  __A = _mm_div_ss(__A, __B);
-  return __builtin_ia32_selectss_128(__U, __A, _mm_setzero_ps());
-}
-
-#define _mm_div_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_divss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_div_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_divss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm_maskz_div_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_divss_round_mask((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4sf)_mm_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_div_sd(__m128d __W, __mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_div_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, __W);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_div_sd(__mmask8 __U,__m128d __A, __m128d __B) {
-  __A = _mm_div_sd(__A, __B);
-  return __builtin_ia32_selectsd_128(__U, __A, _mm_setzero_pd());
-}
-
-#define _mm_div_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_divsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm_mask_div_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_divsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)(__m128d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_div_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_divsd_round_mask((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2df)_mm_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_div_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)((__v8df)__a/(__v8df)__b);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_div_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_div_pd(__A, __B),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_div_pd(__mmask8 __U, __m512d __A, __m512d __B) {
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_div_pd(__A, __B),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_div_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)((__v16sf)__a/(__v16sf)__b);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_div_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_div_ps(__A, __B),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_div_ps(__mmask16 __U, __m512 __A, __m512 __B) {
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_div_ps(__A, __B),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-#define _mm512_div_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_divpd512((__v8df)(__m512d)(A), \
-                                   (__v8df)(__m512d)(B), (int)(R))
-
-#define _mm512_mask_div_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_div_round_pd((A), (B), (R)), \
-                                   (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_div_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                   (__v8df)_mm512_div_round_pd((A), (B), (R)), \
-                                   (__v8df)_mm512_setzero_pd())
-
-#define _mm512_div_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_divps512((__v16sf)(__m512)(A), \
-                                  (__v16sf)(__m512)(B), (int)(R))
-
-#define _mm512_mask_div_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_div_round_ps((A), (B), (R)), \
-                                  (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_div_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                  (__v16sf)_mm512_div_round_ps((A), (B), (R)), \
-                                  (__v16sf)_mm512_setzero_ps())
-
-#define _mm512_roundscale_ps(A, B) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(A), (int)(B), \
-                                         (__v16sf)_mm512_undefined_ps(), \
-                                         (__mmask16)-1, \
-                                         _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_roundscale_ps(A, B, C, imm) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(C), (int)(imm), \
-                                         (__v16sf)(__m512)(A), (__mmask16)(B), \
-                                         _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_roundscale_ps(A, B, imm) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(B), (int)(imm), \
-                                         (__v16sf)_mm512_setzero_ps(), \
-                                         (__mmask16)(A), \
-                                         _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_roundscale_round_ps(A, B, C, imm, R) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(C), (int)(imm), \
-                                         (__v16sf)(__m512)(A), (__mmask16)(B), \
-                                         (int)(R))
-
-#define _mm512_maskz_roundscale_round_ps(A, B, imm, R) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(B), (int)(imm), \
-                                         (__v16sf)_mm512_setzero_ps(), \
-                                         (__mmask16)(A), (int)(R))
-
-#define _mm512_roundscale_round_ps(A, imm, R) \
-  (__m512)__builtin_ia32_rndscaleps_mask((__v16sf)(__m512)(A), (int)(imm), \
-                                         (__v16sf)_mm512_undefined_ps(), \
-                                         (__mmask16)-1, (int)(R))
-
-#define _mm512_roundscale_pd(A, B) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(A), (int)(B), \
-                                          (__v8df)_mm512_undefined_pd(), \
-                                          (__mmask8)-1, \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_roundscale_pd(A, B, C, imm) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(C), (int)(imm), \
-                                          (__v8df)(__m512d)(A), (__mmask8)(B), \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_roundscale_pd(A, B, imm) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(B), (int)(imm), \
-                                          (__v8df)_mm512_setzero_pd(), \
-                                          (__mmask8)(A), \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_roundscale_round_pd(A, B, C, imm, R) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(C), (int)(imm), \
-                                          (__v8df)(__m512d)(A), (__mmask8)(B), \
-                                          (int)(R))
-
-#define _mm512_maskz_roundscale_round_pd(A, B, imm, R) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(B), (int)(imm), \
-                                          (__v8df)_mm512_setzero_pd(), \
-                                          (__mmask8)(A), (int)(R))
-
-#define _mm512_roundscale_round_pd(A, imm, R) \
-  (__m512d)__builtin_ia32_rndscalepd_mask((__v8df)(__m512d)(A), (int)(imm), \
-                                          (__v8df)_mm512_undefined_pd(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm512_fmadd_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)(__m512d)(C), \
-                                           (__mmask8)-1, (int)(R))
-
-
-#define _mm512_mask_fmadd_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)(__m512d)(C), \
-                                           (__mmask8)(U), (int)(R))
-
-
-#define _mm512_mask3_fmadd_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask3((__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-#define _mm512_maskz_fmadd_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_maskz((__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-#define _mm512_fmsub_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           -(__v8df)(__m512d)(C), \
-                                           (__mmask8)-1, (int)(R))
-
-
-#define _mm512_mask_fmsub_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           -(__v8df)(__m512d)(C), \
-                                           (__mmask8)(U), (int)(R))
-
-
-#define _mm512_maskz_fmsub_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_maskz((__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            -(__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-#define _mm512_fnmadd_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask(-(__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)(__m512d)(C), \
-                                           (__mmask8)-1, (int)(R))
-
-
-#define _mm512_mask3_fnmadd_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask3(-(__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-#define _mm512_maskz_fnmadd_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_maskz(-(__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-#define _mm512_fnmsub_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask(-(__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           -(__v8df)(__m512d)(C), \
-                                           (__mmask8)-1, (int)(R))
-
-
-#define _mm512_maskz_fnmsub_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_maskz(-(__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            -(__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fmadd_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    (__v8df) __B,
-                                                    (__v8df) __C,
-                                                    (__mmask8) -1,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fmadd_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    (__v8df) __B,
-                                                    (__v8df) __C,
-                                                    (__mmask8) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmadd_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask3 ((__v8df) __A,
-                                                     (__v8df) __B,
-                                                     (__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmadd_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_maskz ((__v8df) __A,
-                                                     (__v8df) __B,
-                                                     (__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fmsub_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    (__v8df) __B,
-                                                    -(__v8df) __C,
-                                                    (__mmask8) -1,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fmsub_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    (__v8df) __B,
-                                                    -(__v8df) __C,
-                                                    (__mmask8) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmsub_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_maskz ((__v8df) __A,
-                                                     (__v8df) __B,
-                                                     -(__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fnmadd_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    -(__v8df) __B,
-                                                    (__v8df) __C,
-                                                    (__mmask8) -1,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fnmadd_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask3 (-(__v8df) __A,
-                                                     (__v8df) __B,
-                                                     (__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fnmadd_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_maskz (-(__v8df) __A,
-                                                     (__v8df) __B,
-                                                     (__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fnmsub_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    -(__v8df) __B,
-                                                    -(__v8df) __C,
-                                                    (__mmask8) -1,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fnmsub_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_maskz (-(__v8df) __A,
-                                                     (__v8df) __B,
-                                                     -(__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_fmadd_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          (__v16sf)(__m512)(C), \
-                                          (__mmask16)-1, (int)(R))
-
-
-#define _mm512_mask_fmadd_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          (__v16sf)(__m512)(C), \
-                                          (__mmask16)(U), (int)(R))
-
-
-#define _mm512_mask3_fmadd_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask3((__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-#define _mm512_maskz_fmadd_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_maskz((__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-#define _mm512_fmsub_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          -(__v16sf)(__m512)(C), \
-                                          (__mmask16)-1, (int)(R))
-
-
-#define _mm512_mask_fmsub_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          -(__v16sf)(__m512)(C), \
-                                          (__mmask16)(U), (int)(R))
-
-
-#define _mm512_maskz_fmsub_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_maskz((__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           -(__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-#define _mm512_fnmadd_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          -(__v16sf)(__m512)(B), \
-                                          (__v16sf)(__m512)(C), \
-                                          (__mmask16)-1, (int)(R))
-
-
-#define _mm512_mask3_fnmadd_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask3(-(__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-#define _mm512_maskz_fnmadd_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_maskz(-(__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-#define _mm512_fnmsub_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          -(__v16sf)(__m512)(B), \
-                                          -(__v16sf)(__m512)(C), \
-                                          (__mmask16)-1, (int)(R))
-
-
-#define _mm512_maskz_fnmsub_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_maskz(-(__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           -(__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fmadd_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   (__v16sf) __B,
-                                                   (__v16sf) __C,
-                                                   (__mmask16) -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fmadd_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   (__v16sf) __B,
-                                                   (__v16sf) __C,
-                                                   (__mmask16) __U,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmadd_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask3 ((__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    (__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmadd_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_maskz ((__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    (__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fmsub_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   (__v16sf) __B,
-                                                   -(__v16sf) __C,
-                                                   (__mmask16) -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fmsub_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   (__v16sf) __B,
-                                                   -(__v16sf) __C,
-                                                   (__mmask16) __U,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmsub_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_maskz ((__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    -(__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fnmadd_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   -(__v16sf) __B,
-                                                   (__v16sf) __C,
-                                                   (__mmask16) -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fnmadd_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask3 (-(__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    (__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fnmadd_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_maskz (-(__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    (__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fnmsub_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   -(__v16sf) __B,
-                                                   -(__v16sf) __C,
-                                                   (__mmask16) -1,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fnmsub_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_maskz (-(__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    -(__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_fmaddsub_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_mask((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              (__v8df)(__m512d)(C), \
-                                              (__mmask8)-1, (int)(R))
-
-
-#define _mm512_mask_fmaddsub_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_mask((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              (__v8df)(__m512d)(C), \
-                                              (__mmask8)(U), (int)(R))
-
-
-#define _mm512_mask3_fmaddsub_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_mask3((__v8df)(__m512d)(A), \
-                                               (__v8df)(__m512d)(B), \
-                                               (__v8df)(__m512d)(C), \
-                                               (__mmask8)(U), (int)(R))
-
-
-#define _mm512_maskz_fmaddsub_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_maskz((__v8df)(__m512d)(A), \
-                                               (__v8df)(__m512d)(B), \
-                                               (__v8df)(__m512d)(C), \
-                                               (__mmask8)(U), (int)(R))
-
-
-#define _mm512_fmsubadd_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_mask((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              -(__v8df)(__m512d)(C), \
-                                              (__mmask8)-1, (int)(R))
-
-
-#define _mm512_mask_fmsubadd_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_mask((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              -(__v8df)(__m512d)(C), \
-                                              (__mmask8)(U), (int)(R))
-
-
-#define _mm512_maskz_fmsubadd_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddsubpd512_maskz((__v8df)(__m512d)(A), \
-                                               (__v8df)(__m512d)(B), \
-                                               -(__v8df)(__m512d)(C), \
-                                               (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fmaddsub_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_mask ((__v8df) __A,
-                                                      (__v8df) __B,
-                                                      (__v8df) __C,
-                                                      (__mmask8) -1,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fmaddsub_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_mask ((__v8df) __A,
-                                                      (__v8df) __B,
-                                                      (__v8df) __C,
-                                                      (__mmask8) __U,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmaddsub_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_mask3 ((__v8df) __A,
-                                                       (__v8df) __B,
-                                                       (__v8df) __C,
-                                                       (__mmask8) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmaddsub_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_maskz ((__v8df) __A,
-                                                       (__v8df) __B,
-                                                       (__v8df) __C,
-                                                       (__mmask8) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_fmsubadd_pd(__m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_mask ((__v8df) __A,
-                                                       (__v8df) __B,
-                                                       -(__v8df) __C,
-                                                       (__mmask8) -1,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fmsubadd_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_mask ((__v8df) __A,
-                                                       (__v8df) __B,
-                                                       -(__v8df) __C,
-                                                       (__mmask8) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmsubadd_pd(__mmask8 __U, __m512d __A, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddsubpd512_maskz ((__v8df) __A,
-                                                        (__v8df) __B,
-                                                        -(__v8df) __C,
-                                                        (__mmask8) __U,
-                                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_fmaddsub_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_mask((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             (__v16sf)(__m512)(C), \
-                                             (__mmask16)-1, (int)(R))
-
-
-#define _mm512_mask_fmaddsub_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_mask((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             (__v16sf)(__m512)(C), \
-                                             (__mmask16)(U), (int)(R))
-
-
-#define _mm512_mask3_fmaddsub_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_mask3((__v16sf)(__m512)(A), \
-                                              (__v16sf)(__m512)(B), \
-                                              (__v16sf)(__m512)(C), \
-                                              (__mmask16)(U), (int)(R))
-
-
-#define _mm512_maskz_fmaddsub_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_maskz((__v16sf)(__m512)(A), \
-                                              (__v16sf)(__m512)(B), \
-                                              (__v16sf)(__m512)(C), \
-                                              (__mmask16)(U), (int)(R))
-
-
-#define _mm512_fmsubadd_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_mask((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             -(__v16sf)(__m512)(C), \
-                                             (__mmask16)-1, (int)(R))
-
-
-#define _mm512_mask_fmsubadd_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_mask((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             -(__v16sf)(__m512)(C), \
-                                             (__mmask16)(U), (int)(R))
-
-
-#define _mm512_maskz_fmsubadd_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddsubps512_maskz((__v16sf)(__m512)(A), \
-                                              (__v16sf)(__m512)(B), \
-                                              -(__v16sf)(__m512)(C), \
-                                              (__mmask16)(U), (int)(R))
-
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fmaddsub_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_mask ((__v16sf) __A,
-                                                      (__v16sf) __B,
-                                                      (__v16sf) __C,
-                                                      (__mmask16) -1,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fmaddsub_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_mask ((__v16sf) __A,
-                                                      (__v16sf) __B,
-                                                      (__v16sf) __C,
-                                                      (__mmask16) __U,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmaddsub_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_mask3 ((__v16sf) __A,
-                                                       (__v16sf) __B,
-                                                       (__v16sf) __C,
-                                                       (__mmask16) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmaddsub_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_maskz ((__v16sf) __A,
-                                                       (__v16sf) __B,
-                                                       (__v16sf) __C,
-                                                       (__mmask16) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_fmsubadd_ps(__m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_mask ((__v16sf) __A,
-                                                      (__v16sf) __B,
-                                                      -(__v16sf) __C,
-                                                      (__mmask16) -1,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fmsubadd_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_mask ((__v16sf) __A,
-                                                      (__v16sf) __B,
-                                                      -(__v16sf) __C,
-                                                      (__mmask16) __U,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_fmsubadd_ps(__mmask16 __U, __m512 __A, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddsubps512_maskz ((__v16sf) __A,
-                                                       (__v16sf) __B,
-                                                       -(__v16sf) __C,
-                                                       (__mmask16) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask3_fmsub_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmsubpd512_mask3((__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmsub_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d)__builtin_ia32_vfmsubpd512_mask3 ((__v8df) __A,
-                                                    (__v8df) __B,
-                                                    (__v8df) __C,
-                                                    (__mmask8) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask3_fmsub_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmsubps512_mask3((__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmsub_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512)__builtin_ia32_vfmsubps512_mask3 ((__v16sf) __A,
-                                                   (__v16sf) __B,
-                                                   (__v16sf) __C,
-                                                   (__mmask16) __U,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask3_fmsubadd_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmsubaddpd512_mask3((__v8df)(__m512d)(A), \
-                                               (__v8df)(__m512d)(B), \
-                                               (__v8df)(__m512d)(C), \
-                                               (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmsubadd_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d)__builtin_ia32_vfmsubaddpd512_mask3 ((__v8df) __A,
-                                                       (__v8df) __B,
-                                                       (__v8df) __C,
-                                                       (__mmask8) __U,
-                                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask3_fmsubadd_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmsubaddps512_mask3((__v16sf)(__m512)(A), \
-                                              (__v16sf)(__m512)(B), \
-                                              (__v16sf)(__m512)(C), \
-                                              (__mmask16)(U), (int)(R))
-
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fmsubadd_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512)__builtin_ia32_vfmsubaddps512_mask3 ((__v16sf) __A,
-                                                      (__v16sf) __B,
-                                                      (__v16sf) __C,
-                                                      (__mmask16) __U,
-                                                      _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask_fnmadd_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           -(__v8df)(__m512d)(B), \
-                                           (__v8df)(__m512d)(C), \
-                                           (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fnmadd_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    -(__v8df) __B,
-                                                    (__v8df) __C,
-                                                    (__mmask8) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask_fnmadd_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          -(__v16sf)(__m512)(B), \
-                                          (__v16sf)(__m512)(C), \
-                                          (__mmask16)(U), (int)(R))
-
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fnmadd_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   -(__v16sf) __B,
-                                                   (__v16sf) __C,
-                                                   (__mmask16) __U,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask_fnmsub_round_pd(A, U, B, C, R) \
-  (__m512d)__builtin_ia32_vfmaddpd512_mask((__v8df)(__m512d)(A), \
-                                           -(__v8df)(__m512d)(B), \
-                                           -(__v8df)(__m512d)(C), \
-                                           (__mmask8)(U), (int)(R))
-
-
-#define _mm512_mask3_fnmsub_round_pd(A, B, C, U, R) \
-  (__m512d)__builtin_ia32_vfmsubpd512_mask3(-(__v8df)(__m512d)(A), \
-                                            (__v8df)(__m512d)(B), \
-                                            (__v8df)(__m512d)(C), \
-                                            (__mmask8)(U), (int)(R))
-
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_fnmsub_pd(__m512d __A, __mmask8 __U, __m512d __B, __m512d __C)
-{
-  return (__m512d) __builtin_ia32_vfmaddpd512_mask ((__v8df) __A,
-                                                    -(__v8df) __B,
-                                                    -(__v8df) __C,
-                                                    (__mmask8) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask3_fnmsub_pd(__m512d __A, __m512d __B, __m512d __C, __mmask8 __U)
-{
-  return (__m512d) __builtin_ia32_vfmsubpd512_mask3 (-(__v8df) __A,
-                                                     (__v8df) __B,
-                                                     (__v8df) __C,
-                                                     (__mmask8) __U,
-                                                     _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_mask_fnmsub_round_ps(A, U, B, C, R) \
-  (__m512)__builtin_ia32_vfmaddps512_mask((__v16sf)(__m512)(A), \
-                                          -(__v16sf)(__m512)(B), \
-                                          -(__v16sf)(__m512)(C), \
-                                          (__mmask16)(U), (int)(R))
-
-
-#define _mm512_mask3_fnmsub_round_ps(A, B, C, U, R) \
-  (__m512)__builtin_ia32_vfmsubps512_mask3(-(__v16sf)(__m512)(A), \
-                                           (__v16sf)(__m512)(B), \
-                                           (__v16sf)(__m512)(C), \
-                                           (__mmask16)(U), (int)(R))
-
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_fnmsub_ps(__m512 __A, __mmask16 __U, __m512 __B, __m512 __C)
-{
-  return (__m512) __builtin_ia32_vfmaddps512_mask ((__v16sf) __A,
-                                                   -(__v16sf) __B,
-                                                   -(__v16sf) __C,
-                                                   (__mmask16) __U,
-                                                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask3_fnmsub_ps(__m512 __A, __m512 __B, __m512 __C, __mmask16 __U)
-{
-  return (__m512) __builtin_ia32_vfmsubps512_mask3 (-(__v16sf) __A,
-                                                    (__v16sf) __B,
-                                                    (__v16sf) __C,
-                                                    (__mmask16) __U,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-
-
-/* Vector permutations */
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_permutex2var_epi32(__m512i __A, __m512i __I, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_vpermi2vard512((__v16si)__A, (__v16si) __I,
-                                                (__v16si) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_permutex2var_epi32(__m512i __A, __mmask16 __U, __m512i __I,
-                               __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                              (__v16si)_mm512_permutex2var_epi32(__A, __I, __B),
-                              (__v16si)__A);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask2_permutex2var_epi32(__m512i __A, __m512i __I, __mmask16 __U,
-                                __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                              (__v16si)_mm512_permutex2var_epi32(__A, __I, __B),
-                              (__v16si)__I);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutex2var_epi32(__mmask16 __U, __m512i __A, __m512i __I,
-                                __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                              (__v16si)_mm512_permutex2var_epi32(__A, __I, __B),
-                              (__v16si)_mm512_setzero_si512());
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_permutex2var_epi64(__m512i __A, __m512i __I, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_vpermi2varq512((__v8di)__A, (__v8di) __I,
-                                                (__v8di) __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_permutex2var_epi64(__m512i __A, __mmask8 __U, __m512i __I,
-                               __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                               (__v8di)_mm512_permutex2var_epi64(__A, __I, __B),
-                               (__v8di)__A);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask2_permutex2var_epi64(__m512i __A, __m512i __I, __mmask8 __U,
-                                __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                               (__v8di)_mm512_permutex2var_epi64(__A, __I, __B),
-                               (__v8di)__I);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutex2var_epi64(__mmask8 __U, __m512i __A, __m512i __I,
-                                __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                               (__v8di)_mm512_permutex2var_epi64(__A, __I, __B),
-                               (__v8di)_mm512_setzero_si512());
-}
-
-#define _mm512_alignr_epi64(A, B, I) \
-  (__m512i)__builtin_ia32_alignq512((__v8di)(__m512i)(A), \
-                                    (__v8di)(__m512i)(B), (int)(I))
-
-#define _mm512_mask_alignr_epi64(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                 (__v8di)_mm512_alignr_epi64((A), (B), (imm)), \
-                                 (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_alignr_epi64(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                 (__v8di)_mm512_alignr_epi64((A), (B), (imm)), \
-                                 (__v8di)_mm512_setzero_si512())
-
-#define _mm512_alignr_epi32(A, B, I) \
-  (__m512i)__builtin_ia32_alignd512((__v16si)(__m512i)(A), \
-                                    (__v16si)(__m512i)(B), (int)(I))
-
-#define _mm512_mask_alignr_epi32(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                (__v16si)_mm512_alignr_epi32((A), (B), (imm)), \
-                                (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_alignr_epi32(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                (__v16si)_mm512_alignr_epi32((A), (B), (imm)), \
-                                (__v16si)_mm512_setzero_si512())
-/* Vector Extract */
-
-#define _mm512_extractf64x4_pd(A, I) \
-  (__m256d)__builtin_ia32_extractf64x4_mask((__v8df)(__m512d)(A), (int)(I), \
-                                            (__v4df)_mm256_undefined_pd(), \
-                                            (__mmask8)-1)
-
-#define _mm512_mask_extractf64x4_pd(W, U, A, imm) \
-  (__m256d)__builtin_ia32_extractf64x4_mask((__v8df)(__m512d)(A), (int)(imm), \
-                                            (__v4df)(__m256d)(W), \
-                                            (__mmask8)(U))
-
-#define _mm512_maskz_extractf64x4_pd(U, A, imm) \
-  (__m256d)__builtin_ia32_extractf64x4_mask((__v8df)(__m512d)(A), (int)(imm), \
-                                            (__v4df)_mm256_setzero_pd(), \
-                                            (__mmask8)(U))
-
-#define _mm512_extractf32x4_ps(A, I) \
-  (__m128)__builtin_ia32_extractf32x4_mask((__v16sf)(__m512)(A), (int)(I), \
-                                           (__v4sf)_mm_undefined_ps(), \
-                                           (__mmask8)-1)
-
-#define _mm512_mask_extractf32x4_ps(W, U, A, imm) \
-  (__m128)__builtin_ia32_extractf32x4_mask((__v16sf)(__m512)(A), (int)(imm), \
-                                           (__v4sf)(__m128)(W), \
-                                           (__mmask8)(U))
-
-#define _mm512_maskz_extractf32x4_ps(U, A, imm) \
-  (__m128)__builtin_ia32_extractf32x4_mask((__v16sf)(__m512)(A), (int)(imm), \
-                                           (__v4sf)_mm_setzero_ps(), \
-                                           (__mmask8)(U))
-
-/* Vector Blend */
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_blend_pd(__mmask8 __U, __m512d __A, __m512d __W)
-{
-  return (__m512d) __builtin_ia32_selectpd_512 ((__mmask8) __U,
-                 (__v8df) __W,
-                 (__v8df) __A);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_blend_ps(__mmask16 __U, __m512 __A, __m512 __W)
-{
-  return (__m512) __builtin_ia32_selectps_512 ((__mmask16) __U,
-                (__v16sf) __W,
-                (__v16sf) __A);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_blend_epi64(__mmask8 __U, __m512i __A, __m512i __W)
-{
-  return (__m512i) __builtin_ia32_selectq_512 ((__mmask8) __U,
-                (__v8di) __W,
-                (__v8di) __A);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_blend_epi32(__mmask16 __U, __m512i __A, __m512i __W)
-{
-  return (__m512i) __builtin_ia32_selectd_512 ((__mmask16) __U,
-                (__v16si) __W,
-                (__v16si) __A);
-}
-
-/* Compare */
-
-#define _mm512_cmp_round_ps_mask(A, B, P, R) \
-  (__mmask16)__builtin_ia32_cmpps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), (int)(P), \
-                                          (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cmp_round_ps_mask(U, A, B, P, R) \
-  (__mmask16)__builtin_ia32_cmpps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), (int)(P), \
-                                          (__mmask16)(U), (int)(R))
-
-#define _mm512_cmp_ps_mask(A, B, P) \
-  _mm512_cmp_round_ps_mask((A), (B), (P), _MM_FROUND_CUR_DIRECTION)
-#define _mm512_mask_cmp_ps_mask(U, A, B, P) \
-  _mm512_mask_cmp_round_ps_mask((U), (A), (B), (P), _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_cmpeq_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_EQ_OQ)
-#define _mm512_mask_cmpeq_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_EQ_OQ)
-
-#define _mm512_cmplt_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_LT_OS)
-#define _mm512_mask_cmplt_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_LT_OS)
-
-#define _mm512_cmple_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_LE_OS)
-#define _mm512_mask_cmple_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_LE_OS)
-
-#define _mm512_cmpunord_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_UNORD_Q)
-#define _mm512_mask_cmpunord_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_UNORD_Q)
-
-#define _mm512_cmpneq_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_NEQ_UQ)
-#define _mm512_mask_cmpneq_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_NEQ_UQ)
-
-#define _mm512_cmpnlt_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_NLT_US)
-#define _mm512_mask_cmpnlt_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_NLT_US)
-
-#define _mm512_cmpnle_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_NLE_US)
-#define _mm512_mask_cmpnle_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_NLE_US)
-
-#define _mm512_cmpord_ps_mask(A, B) \
-    _mm512_cmp_ps_mask((A), (B), _CMP_ORD_Q)
-#define _mm512_mask_cmpord_ps_mask(k, A, B) \
-    _mm512_mask_cmp_ps_mask((k), (A), (B), _CMP_ORD_Q)
-
-#define _mm512_cmp_round_pd_mask(A, B, P, R) \
-  (__mmask8)__builtin_ia32_cmppd512_mask((__v8df)(__m512d)(A), \
-                                         (__v8df)(__m512d)(B), (int)(P), \
-                                         (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cmp_round_pd_mask(U, A, B, P, R) \
-  (__mmask8)__builtin_ia32_cmppd512_mask((__v8df)(__m512d)(A), \
-                                         (__v8df)(__m512d)(B), (int)(P), \
-                                         (__mmask8)(U), (int)(R))
-
-#define _mm512_cmp_pd_mask(A, B, P) \
-  _mm512_cmp_round_pd_mask((A), (B), (P), _MM_FROUND_CUR_DIRECTION)
-#define _mm512_mask_cmp_pd_mask(U, A, B, P) \
-  _mm512_mask_cmp_round_pd_mask((U), (A), (B), (P), _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_cmpeq_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_EQ_OQ)
-#define _mm512_mask_cmpeq_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_EQ_OQ)
-
-#define _mm512_cmplt_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_LT_OS)
-#define _mm512_mask_cmplt_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_LT_OS)
-
-#define _mm512_cmple_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_LE_OS)
-#define _mm512_mask_cmple_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_LE_OS)
-
-#define _mm512_cmpunord_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_UNORD_Q)
-#define _mm512_mask_cmpunord_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_UNORD_Q)
-
-#define _mm512_cmpneq_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_NEQ_UQ)
-#define _mm512_mask_cmpneq_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_NEQ_UQ)
-
-#define _mm512_cmpnlt_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_NLT_US)
-#define _mm512_mask_cmpnlt_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_NLT_US)
-
-#define _mm512_cmpnle_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_NLE_US)
-#define _mm512_mask_cmpnle_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_NLE_US)
-
-#define _mm512_cmpord_pd_mask(A, B) \
-    _mm512_cmp_pd_mask((A), (B), _CMP_ORD_Q)
-#define _mm512_mask_cmpord_pd_mask(k, A, B) \
-    _mm512_mask_cmp_pd_mask((k), (A), (B), _CMP_ORD_Q)
-
-/* Conversion */
-
-#define _mm512_cvtt_roundps_epu32(A, R) \
-  (__m512i)__builtin_ia32_cvttps2udq512_mask((__v16sf)(__m512)(A), \
-                                             (__v16si)_mm512_undefined_epi32(), \
-                                             (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvtt_roundps_epu32(W, U, A, R) \
-  (__m512i)__builtin_ia32_cvttps2udq512_mask((__v16sf)(__m512)(A), \
-                                             (__v16si)(__m512i)(W), \
-                                             (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvtt_roundps_epu32(U, A, R) \
-  (__m512i)__builtin_ia32_cvttps2udq512_mask((__v16sf)(__m512)(A), \
-                                             (__v16si)_mm512_setzero_si512(), \
-                                             (__mmask16)(U), (int)(R))
-
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvttps_epu32(__m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvttps2udq512_mask ((__v16sf) __A,
-                  (__v16si)
-                  _mm512_setzero_si512 (),
-                  (__mmask16) -1,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvttps_epu32 (__m512i __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvttps2udq512_mask ((__v16sf) __A,
-                   (__v16si) __W,
-                   (__mmask16) __U,
-                   _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvttps_epu32 (__mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvttps2udq512_mask ((__v16sf) __A,
-                   (__v16si) _mm512_setzero_si512 (),
-                   (__mmask16) __U,
-                   _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvt_roundepi32_ps(A, R) \
-  (__m512)__builtin_ia32_cvtdq2ps512_mask((__v16si)(__m512i)(A), \
-                                          (__v16sf)_mm512_setzero_ps(), \
-                                          (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundepi32_ps(W, U, A, R) \
-  (__m512)__builtin_ia32_cvtdq2ps512_mask((__v16si)(__m512i)(A), \
-                                          (__v16sf)(__m512)(W), \
-                                          (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundepi32_ps(U, A, R) \
-  (__m512)__builtin_ia32_cvtdq2ps512_mask((__v16si)(__m512i)(A), \
-                                          (__v16sf)_mm512_setzero_ps(), \
-                                          (__mmask16)(U), (int)(R))
-
-#define _mm512_cvt_roundepu32_ps(A, R) \
-  (__m512)__builtin_ia32_cvtudq2ps512_mask((__v16si)(__m512i)(A), \
-                                           (__v16sf)_mm512_setzero_ps(), \
-                                           (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundepu32_ps(W, U, A, R) \
-  (__m512)__builtin_ia32_cvtudq2ps512_mask((__v16si)(__m512i)(A), \
-                                           (__v16sf)(__m512)(W), \
-                                           (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundepu32_ps(U, A, R) \
-  (__m512)__builtin_ia32_cvtudq2ps512_mask((__v16si)(__m512i)(A), \
-                                           (__v16sf)_mm512_setzero_ps(), \
-                                           (__mmask16)(U), (int)(R))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_cvtepu32_ps (__m512i __A)
-{
-  return (__m512)__builtin_convertvector((__v16su)__A, __v16sf);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu32_ps (__m512 __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_cvtepu32_ps(__A),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu32_ps (__mmask16 __U, __m512i __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_cvtepu32_ps(__A),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32_pd(__m256i __A)
-{
-  return (__m512d)__builtin_convertvector((__v8si)__A, __v8df);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_pd (__m512d __W, __mmask8 __U, __m256i __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                              (__v8df)_mm512_cvtepi32_pd(__A),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi32_pd (__mmask8 __U, __m256i __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                              (__v8df)_mm512_cvtepi32_pd(__A),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32lo_pd(__m512i __A)
-{
-  return (__m512d) _mm512_cvtepi32_pd(_mm512_castsi512_si256(__A));
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32lo_pd(__m512d __W, __mmask8 __U,__m512i __A)
-{
-  return (__m512d) _mm512_mask_cvtepi32_pd(__W, __U, _mm512_castsi512_si256(__A));
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32_ps (__m512i __A)
-{
-  return (__m512)__builtin_convertvector((__v16si)__A, __v16sf);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_ps (__m512 __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_cvtepi32_ps(__A),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi32_ps (__mmask16 __U, __m512i __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_cvtepi32_ps(__A),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtepu32_pd(__m256i __A)
-{
-  return (__m512d)__builtin_convertvector((__v8su)__A, __v8df);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu32_pd (__m512d __W, __mmask8 __U, __m256i __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                              (__v8df)_mm512_cvtepu32_pd(__A),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu32_pd (__mmask8 __U, __m256i __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                              (__v8df)_mm512_cvtepu32_pd(__A),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtepu32lo_pd(__m512i __A)
-{
-  return (__m512d) _mm512_cvtepu32_pd(_mm512_castsi512_si256(__A));
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu32lo_pd(__m512d __W, __mmask8 __U,__m512i __A)
-{
-  return (__m512d) _mm512_mask_cvtepu32_pd(__W, __U, _mm512_castsi512_si256(__A));
-}
-
-#define _mm512_cvt_roundpd_ps(A, R) \
-  (__m256)__builtin_ia32_cvtpd2ps512_mask((__v8df)(__m512d)(A), \
-                                          (__v8sf)_mm256_setzero_ps(), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundpd_ps(W, U, A, R) \
-  (__m256)__builtin_ia32_cvtpd2ps512_mask((__v8df)(__m512d)(A), \
-                                          (__v8sf)(__m256)(W), (__mmask8)(U), \
-                                          (int)(R))
-
-#define _mm512_maskz_cvt_roundpd_ps(U, A, R) \
-  (__m256)__builtin_ia32_cvtpd2ps512_mask((__v8df)(__m512d)(A), \
-                                          (__v8sf)_mm256_setzero_ps(), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m256 __DEFAULT_FN_ATTRS512
-_mm512_cvtpd_ps (__m512d __A)
-{
-  return (__m256) __builtin_ia32_cvtpd2ps512_mask ((__v8df) __A,
-                (__v8sf) _mm256_undefined_ps (),
-                (__mmask8) -1,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256 __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtpd_ps (__m256 __W, __mmask8 __U, __m512d __A)
-{
-  return (__m256) __builtin_ia32_cvtpd2ps512_mask ((__v8df) __A,
-                (__v8sf) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256 __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtpd_ps (__mmask8 __U, __m512d __A)
-{
-  return (__m256) __builtin_ia32_cvtpd2ps512_mask ((__v8df) __A,
-                (__v8sf) _mm256_setzero_ps (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_cvtpd_pslo (__m512d __A)
-{
-  return (__m512) __builtin_shufflevector((__v8sf) _mm512_cvtpd_ps(__A),
-                (__v8sf) _mm256_setzero_ps (),
-                0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtpd_pslo (__m512 __W, __mmask8 __U,__m512d __A)
-{
-  return (__m512) __builtin_shufflevector (
-                (__v8sf) _mm512_mask_cvtpd_ps (_mm512_castps512_ps256(__W),
-                                               __U, __A),
-                (__v8sf) _mm256_setzero_ps (),
-                0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
-}
-
-#define _mm512_cvt_roundps_ph(A, I) \
-  (__m256i)__builtin_ia32_vcvtps2ph512_mask((__v16sf)(__m512)(A), (int)(I), \
-                                            (__v16hi)_mm256_undefined_si256(), \
-                                            (__mmask16)-1)
-
-#define _mm512_mask_cvt_roundps_ph(U, W, A, I) \
-  (__m256i)__builtin_ia32_vcvtps2ph512_mask((__v16sf)(__m512)(A), (int)(I), \
-                                            (__v16hi)(__m256i)(U), \
-                                            (__mmask16)(W))
-
-#define _mm512_maskz_cvt_roundps_ph(W, A, I) \
-  (__m256i)__builtin_ia32_vcvtps2ph512_mask((__v16sf)(__m512)(A), (int)(I), \
-                                            (__v16hi)_mm256_setzero_si256(), \
-                                            (__mmask16)(W))
-
-#define _mm512_cvtps_ph       _mm512_cvt_roundps_ph
-#define _mm512_mask_cvtps_ph  _mm512_mask_cvt_roundps_ph
-#define _mm512_maskz_cvtps_ph _mm512_maskz_cvt_roundps_ph
-
-#define _mm512_cvt_roundph_ps(A, R) \
-  (__m512)__builtin_ia32_vcvtph2ps512_mask((__v16hi)(__m256i)(A), \
-                                           (__v16sf)_mm512_undefined_ps(), \
-                                           (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundph_ps(W, U, A, R) \
-  (__m512)__builtin_ia32_vcvtph2ps512_mask((__v16hi)(__m256i)(A), \
-                                           (__v16sf)(__m512)(W), \
-                                           (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundph_ps(U, A, R) \
-  (__m512)__builtin_ia32_vcvtph2ps512_mask((__v16hi)(__m256i)(A), \
-                                           (__v16sf)_mm512_setzero_ps(), \
-                                           (__mmask16)(U), (int)(R))
-
-
-static  __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_cvtph_ps(__m256i __A)
-{
-  return (__m512) __builtin_ia32_vcvtph2ps512_mask ((__v16hi) __A,
-                (__v16sf)
-                _mm512_setzero_ps (),
-                (__mmask16) -1,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtph_ps (__m512 __W, __mmask16 __U, __m256i __A)
-{
-  return (__m512) __builtin_ia32_vcvtph2ps512_mask ((__v16hi) __A,
-                 (__v16sf) __W,
-                 (__mmask16) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtph_ps (__mmask16 __U, __m256i __A)
-{
-  return (__m512) __builtin_ia32_vcvtph2ps512_mask ((__v16hi) __A,
-                 (__v16sf) _mm512_setzero_ps (),
-                 (__mmask16) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvtt_roundpd_epi32(A, R) \
-  (__m256i)__builtin_ia32_cvttpd2dq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)_mm256_setzero_si256(), \
-                                            (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvtt_roundpd_epi32(W, U, A, R) \
-  (__m256i)__builtin_ia32_cvttpd2dq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)(__m256i)(W), \
-                                            (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_cvtt_roundpd_epi32(U, A, R) \
-  (__m256i)__builtin_ia32_cvttpd2dq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)_mm256_setzero_si256(), \
-                                            (__mmask8)(U), (int)(R))
-
-static __inline __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvttpd_epi32(__m512d __a)
-{
-  return (__m256i)__builtin_ia32_cvttpd2dq512_mask((__v8df) __a,
-                                                   (__v8si)_mm256_setzero_si256(),
-                                                   (__mmask8) -1,
-                                                    _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvttpd_epi32 (__m256i __W, __mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvttpd2dq512_mask ((__v8df) __A,
-                  (__v8si) __W,
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvttpd_epi32 (__mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvttpd2dq512_mask ((__v8df) __A,
-                  (__v8si) _mm256_setzero_si256 (),
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvtt_roundps_epi32(A, R) \
-  (__m512i)__builtin_ia32_cvttps2dq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)_mm512_setzero_si512(), \
-                                            (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvtt_roundps_epi32(W, U, A, R) \
-  (__m512i)__builtin_ia32_cvttps2dq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)(__m512i)(W), \
-                                            (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvtt_roundps_epi32(U, A, R) \
-  (__m512i)__builtin_ia32_cvttps2dq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)_mm512_setzero_si512(), \
-                                            (__mmask16)(U), (int)(R))
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvttps_epi32(__m512 __a)
-{
-  return (__m512i)
-    __builtin_ia32_cvttps2dq512_mask((__v16sf) __a,
-                                     (__v16si) _mm512_setzero_si512 (),
-                                     (__mmask16) -1, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvttps_epi32 (__m512i __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvttps2dq512_mask ((__v16sf) __A,
-                  (__v16si) __W,
-                  (__mmask16) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvttps_epi32 (__mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvttps2dq512_mask ((__v16sf) __A,
-                  (__v16si) _mm512_setzero_si512 (),
-                  (__mmask16) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvt_roundps_epi32(A, R) \
-  (__m512i)__builtin_ia32_cvtps2dq512_mask((__v16sf)(__m512)(A), \
-                                           (__v16si)_mm512_setzero_si512(), \
-                                           (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundps_epi32(W, U, A, R) \
-  (__m512i)__builtin_ia32_cvtps2dq512_mask((__v16sf)(__m512)(A), \
-                                           (__v16si)(__m512i)(W), \
-                                           (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundps_epi32(U, A, R) \
-  (__m512i)__builtin_ia32_cvtps2dq512_mask((__v16sf)(__m512)(A), \
-                                           (__v16si)_mm512_setzero_si512(), \
-                                           (__mmask16)(U), (int)(R))
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtps_epi32 (__m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2dq512_mask ((__v16sf) __A,
-                 (__v16si) _mm512_undefined_epi32 (),
-                 (__mmask16) -1,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtps_epi32 (__m512i __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2dq512_mask ((__v16sf) __A,
-                 (__v16si) __W,
-                 (__mmask16) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtps_epi32 (__mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2dq512_mask ((__v16sf) __A,
-                 (__v16si)
-                 _mm512_setzero_si512 (),
-                 (__mmask16) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvt_roundpd_epi32(A, R) \
-  (__m256i)__builtin_ia32_cvtpd2dq512_mask((__v8df)(__m512d)(A), \
-                                           (__v8si)_mm256_setzero_si256(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundpd_epi32(W, U, A, R) \
-  (__m256i)__builtin_ia32_cvtpd2dq512_mask((__v8df)(__m512d)(A), \
-                                           (__v8si)(__m256i)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundpd_epi32(U, A, R) \
-  (__m256i)__builtin_ia32_cvtpd2dq512_mask((__v8df)(__m512d)(A), \
-                                           (__v8si)_mm256_setzero_si256(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtpd_epi32 (__m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2dq512_mask ((__v8df) __A,
-                 (__v8si)
-                 _mm256_undefined_si256 (),
-                 (__mmask8) -1,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtpd_epi32 (__m256i __W, __mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2dq512_mask ((__v8df) __A,
-                 (__v8si) __W,
-                 (__mmask8) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtpd_epi32 (__mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2dq512_mask ((__v8df) __A,
-                 (__v8si)
-                 _mm256_setzero_si256 (),
-                 (__mmask8) __U,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvt_roundps_epu32(A, R) \
-  (__m512i)__builtin_ia32_cvtps2udq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)_mm512_setzero_si512(), \
-                                            (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundps_epu32(W, U, A, R) \
-  (__m512i)__builtin_ia32_cvtps2udq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)(__m512i)(W), \
-                                            (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundps_epu32(U, A, R) \
-  (__m512i)__builtin_ia32_cvtps2udq512_mask((__v16sf)(__m512)(A), \
-                                            (__v16si)_mm512_setzero_si512(), \
-                                            (__mmask16)(U), (int)(R))
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtps_epu32 ( __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2udq512_mask ((__v16sf) __A,\
-                  (__v16si)\
-                  _mm512_undefined_epi32 (),
-                  (__mmask16) -1,\
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtps_epu32 (__m512i __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2udq512_mask ((__v16sf) __A,
-                  (__v16si) __W,
-                  (__mmask16) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtps_epu32 ( __mmask16 __U, __m512 __A)
-{
-  return (__m512i) __builtin_ia32_cvtps2udq512_mask ((__v16sf) __A,
-                  (__v16si)
-                  _mm512_setzero_si512 (),
-                  (__mmask16) __U ,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_cvt_roundpd_epu32(A, R) \
-  (__m256i)__builtin_ia32_cvtpd2udq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)_mm256_setzero_si256(), \
-                                            (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundpd_epu32(W, U, A, R) \
-  (__m256i)__builtin_ia32_cvtpd2udq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)(__m256i)(W), \
-                                            (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundpd_epu32(U, A, R) \
-  (__m256i)__builtin_ia32_cvtpd2udq512_mask((__v8df)(__m512d)(A), \
-                                            (__v8si)_mm256_setzero_si256(), \
-                                            (__mmask8)(U), (int)(R))
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtpd_epu32 (__m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2udq512_mask ((__v8df) __A,
-                  (__v8si)
-                  _mm256_undefined_si256 (),
-                  (__mmask8) -1,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtpd_epu32 (__m256i __W, __mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2udq512_mask ((__v8df) __A,
-                  (__v8si) __W,
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtpd_epu32 (__mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvtpd2udq512_mask ((__v8df) __A,
-                  (__v8si)
-                  _mm256_setzero_si256 (),
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_cvtsd_f64(__m512d __a)
-{
-  return __a[0];
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_cvtss_f32(__m512 __a)
-{
-  return __a[0];
-}
-
-/* Unpack and Interleave */
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_unpackhi_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)__builtin_shufflevector((__v8df)__a, (__v8df)__b,
-                                          1, 9, 1+2, 9+2, 1+4, 9+4, 1+6, 9+6);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_unpackhi_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                           (__v8df)_mm512_unpackhi_pd(__A, __B),
-                                           (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpackhi_pd(__mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                           (__v8df)_mm512_unpackhi_pd(__A, __B),
-                                           (__v8df)_mm512_setzero_pd());
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_unpacklo_pd(__m512d __a, __m512d __b)
-{
-  return (__m512d)__builtin_shufflevector((__v8df)__a, (__v8df)__b,
-                                          0, 8, 0+2, 8+2, 0+4, 8+4, 0+6, 8+6);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_unpacklo_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                           (__v8df)_mm512_unpacklo_pd(__A, __B),
-                                           (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpacklo_pd (__mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8) __U,
-                                           (__v8df)_mm512_unpacklo_pd(__A, __B),
-                                           (__v8df)_mm512_setzero_pd());
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_unpackhi_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)__builtin_shufflevector((__v16sf)__a, (__v16sf)__b,
-                                         2,    18,    3,    19,
-                                         2+4,  18+4,  3+4,  19+4,
-                                         2+8,  18+8,  3+8,  19+8,
-                                         2+12, 18+12, 3+12, 19+12);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_unpackhi_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16) __U,
-                                          (__v16sf)_mm512_unpackhi_ps(__A, __B),
-                                          (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpackhi_ps (__mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16) __U,
-                                          (__v16sf)_mm512_unpackhi_ps(__A, __B),
-                                          (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_unpacklo_ps(__m512 __a, __m512 __b)
-{
-  return (__m512)__builtin_shufflevector((__v16sf)__a, (__v16sf)__b,
-                                         0,    16,    1,    17,
-                                         0+4,  16+4,  1+4,  17+4,
-                                         0+8,  16+8,  1+8,  17+8,
-                                         0+12, 16+12, 1+12, 17+12);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_unpacklo_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16) __U,
-                                          (__v16sf)_mm512_unpacklo_ps(__A, __B),
-                                          (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpacklo_ps (__mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16) __U,
-                                          (__v16sf)_mm512_unpacklo_ps(__A, __B),
-                                          (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_unpackhi_epi32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_shufflevector((__v16si)__A, (__v16si)__B,
-                                          2,    18,    3,    19,
-                                          2+4,  18+4,  3+4,  19+4,
-                                          2+8,  18+8,  3+8,  19+8,
-                                          2+12, 18+12, 3+12, 19+12);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_unpackhi_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16) __U,
-                                       (__v16si)_mm512_unpackhi_epi32(__A, __B),
-                                       (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpackhi_epi32(__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16) __U,
-                                       (__v16si)_mm512_unpackhi_epi32(__A, __B),
-                                       (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_unpacklo_epi32(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_shufflevector((__v16si)__A, (__v16si)__B,
-                                          0,    16,    1,    17,
-                                          0+4,  16+4,  1+4,  17+4,
-                                          0+8,  16+8,  1+8,  17+8,
-                                          0+12, 16+12, 1+12, 17+12);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_unpacklo_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16) __U,
-                                       (__v16si)_mm512_unpacklo_epi32(__A, __B),
-                                       (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpacklo_epi32(__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16) __U,
-                                       (__v16si)_mm512_unpacklo_epi32(__A, __B),
-                                       (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_unpackhi_epi64(__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_shufflevector((__v8di)__A, (__v8di)__B,
-                                          1, 9, 1+2, 9+2, 1+4, 9+4, 1+6, 9+6);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_unpackhi_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8) __U,
-                                        (__v8di)_mm512_unpackhi_epi64(__A, __B),
-                                        (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpackhi_epi64(__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8) __U,
-                                        (__v8di)_mm512_unpackhi_epi64(__A, __B),
-                                        (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_unpacklo_epi64 (__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_shufflevector((__v8di)__A, (__v8di)__B,
-                                          0, 8, 0+2, 8+2, 0+4, 8+4, 0+6, 8+6);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_unpacklo_epi64 (__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8) __U,
-                                        (__v8di)_mm512_unpacklo_epi64(__A, __B),
-                                        (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_unpacklo_epi64 (__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8) __U,
-                                        (__v8di)_mm512_unpacklo_epi64(__A, __B),
-                                        (__v8di)_mm512_setzero_si512());
-}
-
-
-/* SIMD load ops */
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_loadu_si512 (void const *__P)
-{
-  struct __loadu_si512 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_si512*)__P)->__v;
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_loadu_epi32 (void const *__P)
-{
-  struct __loadu_epi32 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_epi32*)__P)->__v;
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_loadu_epi32 (__m512i __W, __mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_loaddqusi512_mask ((const int *) __P,
-                  (__v16si) __W,
-                  (__mmask16) __U);
-}
-
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_loadu_epi32(__mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_loaddqusi512_mask ((const int *)__P,
-                                                     (__v16si)
-                                                     _mm512_setzero_si512 (),
-                                                     (__mmask16) __U);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_loadu_epi64 (void const *__P)
-{
-  struct __loadu_epi64 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_epi64*)__P)->__v;
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_loadu_epi64 (__m512i __W, __mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_loaddqudi512_mask ((const long long *) __P,
-                  (__v8di) __W,
-                  (__mmask8) __U);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_loadu_epi64(__mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_loaddqudi512_mask ((const long long *)__P,
-                                                     (__v8di)
-                                                     _mm512_setzero_si512 (),
-                                                     (__mmask8) __U);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_loadu_ps (__m512 __W, __mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_loadups512_mask ((const float *) __P,
-                   (__v16sf) __W,
-                   (__mmask16) __U);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_loadu_ps(__mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_loadups512_mask ((const float *)__P,
-                                                  (__v16sf)
-                                                  _mm512_setzero_ps (),
-                                                  (__mmask16) __U);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_loadu_pd (__m512d __W, __mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_loadupd512_mask ((const double *) __P,
-                (__v8df) __W,
-                (__mmask8) __U);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_loadu_pd(__mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_loadupd512_mask ((const double *)__P,
-                                                   (__v8df)
-                                                   _mm512_setzero_pd (),
-                                                   (__mmask8) __U);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_loadu_pd(void const *__p)
-{
-  struct __loadu_pd {
-    __m512d_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_pd*)__p)->__v;
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_loadu_ps(void const *__p)
-{
-  struct __loadu_ps {
-    __m512_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_ps*)__p)->__v;
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_load_ps(void const *__p)
-{
-  return *(__m512*)__p;
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_load_ps (__m512 __W, __mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_loadaps512_mask ((const __v16sf *) __P,
-                   (__v16sf) __W,
-                   (__mmask16) __U);
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_load_ps(__mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_loadaps512_mask ((const __v16sf *)__P,
-                                                  (__v16sf)
-                                                  _mm512_setzero_ps (),
-                                                  (__mmask16) __U);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_load_pd(void const *__p)
-{
-  return *(__m512d*)__p;
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_load_pd (__m512d __W, __mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_loadapd512_mask ((const __v8df *) __P,
-                          (__v8df) __W,
-                          (__mmask8) __U);
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_load_pd(__mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_loadapd512_mask ((const __v8df *)__P,
-                                                   (__v8df)
-                                                   _mm512_setzero_pd (),
-                                                   (__mmask8) __U);
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_load_si512 (void const *__P)
-{
-  return *(__m512i *) __P;
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_load_epi32 (void const *__P)
-{
-  return *(__m512i *) __P;
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_load_epi64 (void const *__P)
-{
-  return *(__m512i *) __P;
-}
-
-/* SIMD store ops */
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_storeu_epi64 (void *__P, __m512i __A)
-{
-  struct __storeu_epi64 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_epi64*)__P)->__v = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_storeu_epi64(void *__P, __mmask8 __U, __m512i __A)
-{
-  __builtin_ia32_storedqudi512_mask ((long long *)__P, (__v8di) __A,
-                                     (__mmask8) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_storeu_si512 (void *__P, __m512i __A)
-{
-  struct __storeu_si512 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_si512*)__P)->__v = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_storeu_epi32 (void *__P, __m512i __A)
-{
-  struct __storeu_epi32 {
-    __m512i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_epi32*)__P)->__v = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_storeu_epi32(void *__P, __mmask16 __U, __m512i __A)
-{
-  __builtin_ia32_storedqusi512_mask ((int *)__P, (__v16si) __A,
-                                     (__mmask16) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_storeu_pd(void *__P, __mmask8 __U, __m512d __A)
-{
-  __builtin_ia32_storeupd512_mask ((double *)__P, (__v8df) __A, (__mmask8) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_storeu_pd(void *__P, __m512d __A)
-{
-  struct __storeu_pd {
-    __m512d_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_pd*)__P)->__v = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_storeu_ps(void *__P, __mmask16 __U, __m512 __A)
-{
-  __builtin_ia32_storeups512_mask ((float *)__P, (__v16sf) __A,
-                                   (__mmask16) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_storeu_ps(void *__P, __m512 __A)
-{
-  struct __storeu_ps {
-    __m512_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_ps*)__P)->__v = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_store_pd(void *__P, __mmask8 __U, __m512d __A)
-{
-  __builtin_ia32_storeapd512_mask ((__v8df *)__P, (__v8df) __A, (__mmask8) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_store_pd(void *__P, __m512d __A)
-{
-  *(__m512d*)__P = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_mask_store_ps(void *__P, __mmask16 __U, __m512 __A)
-{
-  __builtin_ia32_storeaps512_mask ((__v16sf *)__P, (__v16sf) __A,
-                                   (__mmask16) __U);
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_store_ps(void *__P, __m512 __A)
-{
-  *(__m512*)__P = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_store_si512 (void *__P, __m512i __A)
-{
-  *(__m512i *) __P = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_store_epi32 (void *__P, __m512i __A)
-{
-  *(__m512i *) __P = __A;
-}
-
-static __inline void __DEFAULT_FN_ATTRS512
-_mm512_store_epi64 (void *__P, __m512i __A)
-{
-  *(__m512i *) __P = __A;
-}
-
-/* Mask ops */
-
-static __inline __mmask16 __DEFAULT_FN_ATTRS
-_mm512_knot(__mmask16 __M)
-{
-  return __builtin_ia32_knothi(__M);
-}
-
-/* Integer compare */
-
-#define _mm512_cmpeq_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_EQ)
-#define _mm512_mask_cmpeq_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_EQ)
-#define _mm512_cmpge_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_GE)
-#define _mm512_mask_cmpge_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_GE)
-#define _mm512_cmpgt_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_GT)
-#define _mm512_mask_cmpgt_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_GT)
-#define _mm512_cmple_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_LE)
-#define _mm512_mask_cmple_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_LE)
-#define _mm512_cmplt_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_LT)
-#define _mm512_mask_cmplt_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_LT)
-#define _mm512_cmpneq_epi32_mask(A, B) \
-    _mm512_cmp_epi32_mask((A), (B), _MM_CMPINT_NE)
-#define _mm512_mask_cmpneq_epi32_mask(k, A, B) \
-    _mm512_mask_cmp_epi32_mask((k), (A), (B), _MM_CMPINT_NE)
-
-#define _mm512_cmpeq_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_EQ)
-#define _mm512_mask_cmpeq_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_EQ)
-#define _mm512_cmpge_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_GE)
-#define _mm512_mask_cmpge_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_GE)
-#define _mm512_cmpgt_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_GT)
-#define _mm512_mask_cmpgt_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_GT)
-#define _mm512_cmple_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_LE)
-#define _mm512_mask_cmple_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_LE)
-#define _mm512_cmplt_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_LT)
-#define _mm512_mask_cmplt_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_LT)
-#define _mm512_cmpneq_epu32_mask(A, B) \
-    _mm512_cmp_epu32_mask((A), (B), _MM_CMPINT_NE)
-#define _mm512_mask_cmpneq_epu32_mask(k, A, B) \
-    _mm512_mask_cmp_epu32_mask((k), (A), (B), _MM_CMPINT_NE)
-
-#define _mm512_cmpeq_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_EQ)
-#define _mm512_mask_cmpeq_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_EQ)
-#define _mm512_cmpge_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_GE)
-#define _mm512_mask_cmpge_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_GE)
-#define _mm512_cmpgt_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_GT)
-#define _mm512_mask_cmpgt_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_GT)
-#define _mm512_cmple_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_LE)
-#define _mm512_mask_cmple_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_LE)
-#define _mm512_cmplt_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_LT)
-#define _mm512_mask_cmplt_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_LT)
-#define _mm512_cmpneq_epi64_mask(A, B) \
-    _mm512_cmp_epi64_mask((A), (B), _MM_CMPINT_NE)
-#define _mm512_mask_cmpneq_epi64_mask(k, A, B) \
-    _mm512_mask_cmp_epi64_mask((k), (A), (B), _MM_CMPINT_NE)
-
-#define _mm512_cmpeq_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_EQ)
-#define _mm512_mask_cmpeq_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_EQ)
-#define _mm512_cmpge_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_GE)
-#define _mm512_mask_cmpge_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_GE)
-#define _mm512_cmpgt_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_GT)
-#define _mm512_mask_cmpgt_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_GT)
-#define _mm512_cmple_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_LE)
-#define _mm512_mask_cmple_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_LE)
-#define _mm512_cmplt_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_LT)
-#define _mm512_mask_cmplt_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_LT)
-#define _mm512_cmpneq_epu64_mask(A, B) \
-    _mm512_cmp_epu64_mask((A), (B), _MM_CMPINT_NE)
-#define _mm512_mask_cmpneq_epu64_mask(k, A, B) \
-    _mm512_mask_cmp_epu64_mask((k), (A), (B), _MM_CMPINT_NE)
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi8_epi32(__m128i __A)
-{
-  /* This function always performs a signed extension, but __v16qi is a char
-     which may be signed or unsigned, so use __v16qs. */
-  return (__m512i)__builtin_convertvector((__v16qs)__A, __v16si);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi8_epi32(__m512i __W, __mmask16 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_cvtepi8_epi32(__A),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi8_epi32(__mmask16 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_cvtepi8_epi32(__A),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi8_epi64(__m128i __A)
-{
-  /* This function always performs a signed extension, but __v16qi is a char
-     which may be signed or unsigned, so use __v16qs. */
-  return (__m512i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__A, (__v16qs)__A, 0, 1, 2, 3, 4, 5, 6, 7), __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi8_epi64(__m512i __W, __mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi8_epi64(__A),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi8_epi64(__mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi8_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512 ());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32_epi64(__m256i __X)
-{
-  return (__m512i)__builtin_convertvector((__v8si)__X, __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_epi64(__m512i __W, __mmask8 __U, __m256i __X)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi32_epi64(__X),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi32_epi64(__mmask8 __U, __m256i __X)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi32_epi64(__X),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi16_epi32(__m256i __A)
-{
-  return (__m512i)__builtin_convertvector((__v16hi)__A, __v16si);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi16_epi32(__m512i __W, __mmask16 __U, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                            (__v16si)_mm512_cvtepi16_epi32(__A),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi16_epi32(__mmask16 __U, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                            (__v16si)_mm512_cvtepi16_epi32(__A),
-                                            (__v16si)_mm512_setzero_si512 ());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi16_epi64(__m128i __A)
-{
-  return (__m512i)__builtin_convertvector((__v8hi)__A, __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi16_epi64(__m512i __W, __mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi16_epi64(__A),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi16_epi64(__mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepi16_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepu8_epi32(__m128i __A)
-{
-  return (__m512i)__builtin_convertvector((__v16qu)__A, __v16si);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu8_epi32(__m512i __W, __mmask16 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_cvtepu8_epi32(__A),
-                                             (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu8_epi32(__mmask16 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                             (__v16si)_mm512_cvtepu8_epi32(__A),
-                                             (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepu8_epi64(__m128i __A)
-{
-  return (__m512i)__builtin_convertvector(__builtin_shufflevector((__v16qu)__A, (__v16qu)__A, 0, 1, 2, 3, 4, 5, 6, 7), __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu8_epi64(__m512i __W, __mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu8_epi64(__A),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu8_epi64(__mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu8_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepu32_epi64(__m256i __X)
-{
-  return (__m512i)__builtin_convertvector((__v8su)__X, __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu32_epi64(__m512i __W, __mmask8 __U, __m256i __X)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu32_epi64(__X),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu32_epi64(__mmask8 __U, __m256i __X)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu32_epi64(__X),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepu16_epi32(__m256i __A)
-{
-  return (__m512i)__builtin_convertvector((__v16hu)__A, __v16si);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu16_epi32(__m512i __W, __mmask16 __U, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                            (__v16si)_mm512_cvtepu16_epi32(__A),
-                                            (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu16_epi32(__mmask16 __U, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                            (__v16si)_mm512_cvtepu16_epi32(__A),
-                                            (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_cvtepu16_epi64(__m128i __A)
-{
-  return (__m512i)__builtin_convertvector((__v8hu)__A, __v8di);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepu16_epi64(__m512i __W, __mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu16_epi64(__A),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepu16_epi64(__mmask8 __U, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_cvtepu16_epi64(__A),
-                                             (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_rorv_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_prorvd512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_rorv_epi32 (__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                           (__v16si)_mm512_rorv_epi32(__A, __B),
-                                           (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_rorv_epi32 (__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                           (__v16si)_mm512_rorv_epi32(__A, __B),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_rorv_epi64 (__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_prorvq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_rorv_epi64 (__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                                            (__v8di)_mm512_rorv_epi64(__A, __B),
-                                            (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_rorv_epi64 (__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                                            (__v8di)_mm512_rorv_epi64(__A, __B),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-
-
-#define _mm512_cmp_epi32_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_cmpd512_mask((__v16si)(__m512i)(a), \
-                                         (__v16si)(__m512i)(b), (int)(p), \
-                                         (__mmask16)-1)
-
-#define _mm512_cmp_epu32_mask(a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpd512_mask((__v16si)(__m512i)(a), \
-                                          (__v16si)(__m512i)(b), (int)(p), \
-                                          (__mmask16)-1)
-
-#define _mm512_cmp_epi64_mask(a, b, p) \
-  (__mmask8)__builtin_ia32_cmpq512_mask((__v8di)(__m512i)(a), \
-                                        (__v8di)(__m512i)(b), (int)(p), \
-                                        (__mmask8)-1)
-
-#define _mm512_cmp_epu64_mask(a, b, p) \
-  (__mmask8)__builtin_ia32_ucmpq512_mask((__v8di)(__m512i)(a), \
-                                         (__v8di)(__m512i)(b), (int)(p), \
-                                         (__mmask8)-1)
-
-#define _mm512_mask_cmp_epi32_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_cmpd512_mask((__v16si)(__m512i)(a), \
-                                         (__v16si)(__m512i)(b), (int)(p), \
-                                         (__mmask16)(m))
-
-#define _mm512_mask_cmp_epu32_mask(m, a, b, p) \
-  (__mmask16)__builtin_ia32_ucmpd512_mask((__v16si)(__m512i)(a), \
-                                          (__v16si)(__m512i)(b), (int)(p), \
-                                          (__mmask16)(m))
-
-#define _mm512_mask_cmp_epi64_mask(m, a, b, p) \
-  (__mmask8)__builtin_ia32_cmpq512_mask((__v8di)(__m512i)(a), \
-                                        (__v8di)(__m512i)(b), (int)(p), \
-                                        (__mmask8)(m))
-
-#define _mm512_mask_cmp_epu64_mask(m, a, b, p) \
-  (__mmask8)__builtin_ia32_ucmpq512_mask((__v8di)(__m512i)(a), \
-                                         (__v8di)(__m512i)(b), (int)(p), \
-                                         (__mmask8)(m))
-
-#define _mm512_rol_epi32(a, b) \
-  (__m512i)__builtin_ia32_prold512((__v16si)(__m512i)(a), (int)(b))
-
-#define _mm512_mask_rol_epi32(W, U, a, b) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_rol_epi32((a), (b)), \
-                                      (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_rol_epi32(U, a, b) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_rol_epi32((a), (b)), \
-                                      (__v16si)_mm512_setzero_si512())
-
-#define _mm512_rol_epi64(a, b) \
-  (__m512i)__builtin_ia32_prolq512((__v8di)(__m512i)(a), (int)(b))
-
-#define _mm512_mask_rol_epi64(W, U, a, b) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_rol_epi64((a), (b)), \
-                                      (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_rol_epi64(U, a, b) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_rol_epi64((a), (b)), \
-                                      (__v8di)_mm512_setzero_si512())
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_rolv_epi32 (__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_prolvd512((__v16si)__A, (__v16si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_rolv_epi32 (__m512i __W, __mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                           (__v16si)_mm512_rolv_epi32(__A, __B),
-                                           (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_rolv_epi32 (__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512(__U,
-                                           (__v16si)_mm512_rolv_epi32(__A, __B),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_rolv_epi64 (__m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_prolvq512((__v8di)__A, (__v8di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_rolv_epi64 (__m512i __W, __mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                                            (__v8di)_mm512_rolv_epi64(__A, __B),
-                                            (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_rolv_epi64 (__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512(__U,
-                                            (__v8di)_mm512_rolv_epi64(__A, __B),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-#define _mm512_ror_epi32(A, B) \
-  (__m512i)__builtin_ia32_prord512((__v16si)(__m512i)(A), (int)(B))
-
-#define _mm512_mask_ror_epi32(W, U, A, B) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_ror_epi32((A), (B)), \
-                                      (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_ror_epi32(U, A, B) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_ror_epi32((A), (B)), \
-                                      (__v16si)_mm512_setzero_si512())
-
-#define _mm512_ror_epi64(A, B) \
-  (__m512i)__builtin_ia32_prorq512((__v8di)(__m512i)(A), (int)(B))
-
-#define _mm512_mask_ror_epi64(W, U, A, B) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_ror_epi64((A), (B)), \
-                                      (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_ror_epi64(U, A, B) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_ror_epi64((A), (B)), \
-                                      (__v8di)_mm512_setzero_si512())
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_slli_epi32(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_pslldi512((__v16si)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_slli_epi32(__m512i __W, __mmask16 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_slli_epi32(__A, __B),
-                                         (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_slli_epi32(__mmask16 __U, __m512i __A, int __B) {
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_slli_epi32(__A, __B),
-                                         (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_slli_epi64(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_psllqi512((__v8di)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_slli_epi64(__m512i __W, __mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_slli_epi64(__A, __B),
-                                          (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_slli_epi64(__mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_slli_epi64(__A, __B),
-                                          (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srli_epi32(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_psrldi512((__v16si)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srli_epi32(__m512i __W, __mmask16 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_srli_epi32(__A, __B),
-                                         (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srli_epi32(__mmask16 __U, __m512i __A, int __B) {
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_srli_epi32(__A, __B),
-                                         (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srli_epi64(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_psrlqi512((__v8di)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srli_epi64(__m512i __W, __mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_srli_epi64(__A, __B),
-                                          (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srli_epi64(__mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_srli_epi64(__A, __B),
-                                          (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_load_epi32 (__m512i __W, __mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_movdqa32load512_mask ((const __v16si *) __P,
-              (__v16si) __W,
-              (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_load_epi32 (__mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_movdqa32load512_mask ((const __v16si *) __P,
-              (__v16si)
-              _mm512_setzero_si512 (),
-              (__mmask16) __U);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_store_epi32 (void *__P, __mmask16 __U, __m512i __A)
-{
-  __builtin_ia32_movdqa32store512_mask ((__v16si *) __P, (__v16si) __A,
-          (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mov_epi32 (__m512i __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_selectd_512 ((__mmask16) __U,
-                 (__v16si) __A,
-                 (__v16si) __W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_mov_epi32 (__mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_selectd_512 ((__mmask16) __U,
-                 (__v16si) __A,
-                 (__v16si) _mm512_setzero_si512 ());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_mov_epi64 (__m512i __W, __mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_selectq_512 ((__mmask8) __U,
-                 (__v8di) __A,
-                 (__v8di) __W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_mov_epi64 (__mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_selectq_512 ((__mmask8) __U,
-                 (__v8di) __A,
-                 (__v8di) _mm512_setzero_si512 ());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_load_epi64 (__m512i __W, __mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_movdqa64load512_mask ((const __v8di *) __P,
-              (__v8di) __W,
-              (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_load_epi64 (__mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_movdqa64load512_mask ((const __v8di *) __P,
-              (__v8di)
-              _mm512_setzero_si512 (),
-              (__mmask8) __U);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_store_epi64 (void *__P, __mmask8 __U, __m512i __A)
-{
-  __builtin_ia32_movdqa64store512_mask ((__v8di *) __P, (__v8di) __A,
-          (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_movedup_pd (__m512d __A)
-{
-  return (__m512d)__builtin_shufflevector((__v8df)__A, (__v8df)__A,
-                                          0, 0, 2, 2, 4, 4, 6, 6);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_movedup_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_movedup_pd(__A),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_movedup_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_movedup_pd(__A),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-#define _mm512_fixupimm_round_pd(A, B, C, imm, R) \
-  (__m512d)__builtin_ia32_fixupimmpd512_mask((__v8df)(__m512d)(A), \
-                                             (__v8df)(__m512d)(B), \
-                                             (__v8di)(__m512i)(C), (int)(imm), \
-                                             (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_fixupimm_round_pd(A, U, B, C, imm, R) \
-  (__m512d)__builtin_ia32_fixupimmpd512_mask((__v8df)(__m512d)(A), \
-                                             (__v8df)(__m512d)(B), \
-                                             (__v8di)(__m512i)(C), (int)(imm), \
-                                             (__mmask8)(U), (int)(R))
-
-#define _mm512_fixupimm_pd(A, B, C, imm) \
-  (__m512d)__builtin_ia32_fixupimmpd512_mask((__v8df)(__m512d)(A), \
-                                             (__v8df)(__m512d)(B), \
-                                             (__v8di)(__m512i)(C), (int)(imm), \
-                                             (__mmask8)-1, \
-                                             _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_fixupimm_pd(A, U, B, C, imm) \
-  (__m512d)__builtin_ia32_fixupimmpd512_mask((__v8df)(__m512d)(A), \
-                                             (__v8df)(__m512d)(B), \
-                                             (__v8di)(__m512i)(C), (int)(imm), \
-                                             (__mmask8)(U), \
-                                             _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_fixupimm_round_pd(U, A, B, C, imm, R) \
-  (__m512d)__builtin_ia32_fixupimmpd512_maskz((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              (__v8di)(__m512i)(C), \
-                                              (int)(imm), (__mmask8)(U), \
-                                              (int)(R))
-
-#define _mm512_maskz_fixupimm_pd(U, A, B, C, imm) \
-  (__m512d)__builtin_ia32_fixupimmpd512_maskz((__v8df)(__m512d)(A), \
-                                              (__v8df)(__m512d)(B), \
-                                              (__v8di)(__m512i)(C), \
-                                              (int)(imm), (__mmask8)(U), \
-                                              _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_fixupimm_round_ps(A, B, C, imm, R) \
-  (__m512)__builtin_ia32_fixupimmps512_mask((__v16sf)(__m512)(A), \
-                                            (__v16sf)(__m512)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_fixupimm_round_ps(A, U, B, C, imm, R) \
-  (__m512)__builtin_ia32_fixupimmps512_mask((__v16sf)(__m512)(A), \
-                                            (__v16sf)(__m512)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)(U), (int)(R))
-
-#define _mm512_fixupimm_ps(A, B, C, imm) \
-  (__m512)__builtin_ia32_fixupimmps512_mask((__v16sf)(__m512)(A), \
-                                            (__v16sf)(__m512)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)-1, \
-                                            _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_fixupimm_ps(A, U, B, C, imm) \
-  (__m512)__builtin_ia32_fixupimmps512_mask((__v16sf)(__m512)(A), \
-                                            (__v16sf)(__m512)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)(U), \
-                                            _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_fixupimm_round_ps(U, A, B, C, imm, R) \
-  (__m512)__builtin_ia32_fixupimmps512_maskz((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             (__v16si)(__m512i)(C), \
-                                             (int)(imm), (__mmask16)(U), \
-                                             (int)(R))
-
-#define _mm512_maskz_fixupimm_ps(U, A, B, C, imm) \
-  (__m512)__builtin_ia32_fixupimmps512_maskz((__v16sf)(__m512)(A), \
-                                             (__v16sf)(__m512)(B), \
-                                             (__v16si)(__m512i)(C), \
-                                             (int)(imm), (__mmask16)(U), \
-                                             _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_fixupimm_round_sd(A, B, C, imm, R) \
-  (__m128d)__builtin_ia32_fixupimmsd_mask((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          (__v2di)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)-1, (int)(R))
-
-#define _mm_mask_fixupimm_round_sd(A, U, B, C, imm, R) \
-  (__m128d)__builtin_ia32_fixupimmsd_mask((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          (__v2di)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)(U), (int)(R))
-
-#define _mm_fixupimm_sd(A, B, C, imm) \
-  (__m128d)__builtin_ia32_fixupimmsd_mask((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          (__v2di)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)-1, \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_fixupimm_sd(A, U, B, C, imm) \
-  (__m128d)__builtin_ia32_fixupimmsd_mask((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          (__v2di)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)(U), \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_fixupimm_round_sd(U, A, B, C, imm, R) \
-  (__m128d)__builtin_ia32_fixupimmsd_maskz((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2di)(__m128i)(C), (int)(imm), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_fixupimm_sd(U, A, B, C, imm) \
-  (__m128d)__builtin_ia32_fixupimmsd_maskz((__v2df)(__m128d)(A), \
-                                           (__v2df)(__m128d)(B), \
-                                           (__v2di)(__m128i)(C), (int)(imm), \
-                                           (__mmask8)(U), \
-                                           _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_fixupimm_round_ss(A, B, C, imm, R) \
-  (__m128)__builtin_ia32_fixupimmss_mask((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         (__v4si)(__m128i)(C), (int)(imm), \
-                                         (__mmask8)-1, (int)(R))
-
-#define _mm_mask_fixupimm_round_ss(A, U, B, C, imm, R) \
-  (__m128)__builtin_ia32_fixupimmss_mask((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         (__v4si)(__m128i)(C), (int)(imm), \
-                                         (__mmask8)(U), (int)(R))
-
-#define _mm_fixupimm_ss(A, B, C, imm) \
-  (__m128)__builtin_ia32_fixupimmss_mask((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         (__v4si)(__m128i)(C), (int)(imm), \
-                                         (__mmask8)-1, \
-                                         _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_fixupimm_ss(A, U, B, C, imm) \
-  (__m128)__builtin_ia32_fixupimmss_mask((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         (__v4si)(__m128i)(C), (int)(imm), \
-                                         (__mmask8)(U), \
-                                         _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_fixupimm_round_ss(U, A, B, C, imm, R) \
-  (__m128)__builtin_ia32_fixupimmss_maskz((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4si)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_fixupimm_ss(U, A, B, C, imm) \
-  (__m128)__builtin_ia32_fixupimmss_maskz((__v4sf)(__m128)(A), \
-                                          (__v4sf)(__m128)(B), \
-                                          (__v4si)(__m128i)(C), (int)(imm), \
-                                          (__mmask8)(U), \
-                                          _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_getexp_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_getexpsd128_round_mask((__v2df)(__m128d)(A), \
-                                                 (__v2df)(__m128d)(B), \
-                                                 (__v2df)_mm_setzero_pd(), \
-                                                 (__mmask8)-1, (int)(R))
-
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_getexp_sd (__m128d __A, __m128d __B)
-{
-  return (__m128d) __builtin_ia32_getexpsd128_round_mask ((__v2df) __A,
-                 (__v2df) __B, (__v2df) _mm_setzero_pd(), (__mmask8) -1, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_getexp_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_getexpsd128_round_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) __W,
-          (__mmask8) __U,
-          _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_getexp_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_getexpsd128_round_mask((__v2df)(__m128d)(A), \
-                                                 (__v2df)(__m128d)(B), \
-                                                 (__v2df)(__m128d)(W), \
-                                                 (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_getexp_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_getexpsd128_round_mask ( (__v2df) __A,
-          (__v2df) __B,
-          (__v2df) _mm_setzero_pd (),
-          (__mmask8) __U,
-          _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_getexp_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_getexpsd128_round_mask((__v2df)(__m128d)(A), \
-                                                 (__v2df)(__m128d)(B), \
-                                                 (__v2df)_mm_setzero_pd(), \
-                                                 (__mmask8)(U), (int)(R))
-
-#define _mm_getexp_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_getexpss128_round_mask((__v4sf)(__m128)(A), \
-                                                (__v4sf)(__m128)(B), \
-                                                (__v4sf)_mm_setzero_ps(), \
-                                                (__mmask8)-1, (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_getexp_ss (__m128 __A, __m128 __B)
-{
-  return (__m128) __builtin_ia32_getexpss128_round_mask ((__v4sf) __A,
-                (__v4sf) __B, (__v4sf)  _mm_setzero_ps(), (__mmask8) -1, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_getexp_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_getexpss128_round_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) __W,
-          (__mmask8) __U,
-          _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_getexp_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_getexpss128_round_mask((__v4sf)(__m128)(A), \
-                                                (__v4sf)(__m128)(B), \
-                                                (__v4sf)(__m128)(W), \
-                                                (__mmask8)(U), (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_getexp_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_getexpss128_round_mask ((__v4sf) __A,
-          (__v4sf) __B,
-          (__v4sf) _mm_setzero_ps (),
-          (__mmask8) __U,
-          _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_getexp_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_getexpss128_round_mask((__v4sf)(__m128)(A), \
-                                                (__v4sf)(__m128)(B), \
-                                                (__v4sf)_mm_setzero_ps(), \
-                                                (__mmask8)(U), (int)(R))
-
-#define _mm_getmant_round_sd(A, B, C, D, R) \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)_mm_setzero_pd(), \
-                                               (__mmask8)-1, (int)(R))
-
-#define _mm_getmant_sd(A, B, C, D)  \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)_mm_setzero_pd(), \
-                                               (__mmask8)-1, \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_getmant_sd(W, U, A, B, C, D) \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)(__m128d)(W), \
-                                               (__mmask8)(U), \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_getmant_round_sd(W, U, A, B, C, D, R) \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)(__m128d)(W), \
-                                               (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_getmant_sd(U, A, B, C, D) \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)_mm_setzero_pd(), \
-                                               (__mmask8)(U), \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_getmant_round_sd(U, A, B, C, D, R) \
-  (__m128d)__builtin_ia32_getmantsd_round_mask((__v2df)(__m128d)(A), \
-                                               (__v2df)(__m128d)(B), \
-                                               (int)(((D)<<2) | (C)), \
-                                               (__v2df)_mm_setzero_pd(), \
-                                               (__mmask8)(U), (int)(R))
-
-#define _mm_getmant_round_ss(A, B, C, D, R) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)_mm_setzero_ps(), \
-                                              (__mmask8)-1, (int)(R))
-
-#define _mm_getmant_ss(A, B, C, D) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)_mm_setzero_ps(), \
-                                              (__mmask8)-1, \
-                                              _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_getmant_ss(W, U, A, B, C, D) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)(__m128)(W), \
-                                              (__mmask8)(U), \
-                                              _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_getmant_round_ss(W, U, A, B, C, D, R) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)(__m128)(W), \
-                                              (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_getmant_ss(U, A, B, C, D) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)_mm_setzero_ps(), \
-                                              (__mmask8)(U), \
-                                              _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_getmant_round_ss(U, A, B, C, D, R) \
-  (__m128)__builtin_ia32_getmantss_round_mask((__v4sf)(__m128)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (int)(((D)<<2) | (C)), \
-                                              (__v4sf)_mm_setzero_ps(), \
-                                              (__mmask8)(U), (int)(R))
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kmov (__mmask16 __A)
-{
-  return  __A;
-}
-
-#define _mm_comi_round_sd(A, B, P, R) \
-  (int)__builtin_ia32_vcomisd((__v2df)(__m128d)(A), (__v2df)(__m128d)(B), \
-                              (int)(P), (int)(R))
-
-#define _mm_comi_round_ss(A, B, P, R) \
-  (int)__builtin_ia32_vcomiss((__v4sf)(__m128)(A), (__v4sf)(__m128)(B), \
-                              (int)(P), (int)(R))
-
-#ifdef __x86_64__
-#define _mm_cvt_roundsd_si64(A, R) \
-  (long long)__builtin_ia32_vcvtsd2si64((__v2df)(__m128d)(A), (int)(R))
-#endif
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sll_epi32(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_pslld512((__v16si) __A, (__v4si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sll_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_sll_epi32(__A, __B),
-                                          (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sll_epi32(__mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_sll_epi32(__A, __B),
-                                          (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sll_epi64(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_psllq512((__v8di)__A, (__v2di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sll_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                             (__v8di)_mm512_sll_epi64(__A, __B),
-                                             (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sll_epi64(__mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                           (__v8di)_mm512_sll_epi64(__A, __B),
-                                           (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sllv_epi32(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psllv16si((__v16si)__X, (__v16si)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sllv_epi32(__m512i __W, __mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_sllv_epi32(__X, __Y),
-                                           (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sllv_epi32(__mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_sllv_epi32(__X, __Y),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sllv_epi64(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psllv8di((__v8di)__X, (__v8di)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sllv_epi64(__m512i __W, __mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_sllv_epi64(__X, __Y),
-                                            (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sllv_epi64(__mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_sllv_epi64(__X, __Y),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sra_epi32(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_psrad512((__v16si) __A, (__v4si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sra_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_sra_epi32(__A, __B),
-                                          (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sra_epi32(__mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_sra_epi32(__A, __B),
-                                          (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_sra_epi64(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_psraq512((__v8di)__A, (__v2di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_sra_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                           (__v8di)_mm512_sra_epi64(__A, __B),
-                                           (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_sra_epi64(__mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                           (__v8di)_mm512_sra_epi64(__A, __B),
-                                           (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srav_epi32(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psrav16si((__v16si)__X, (__v16si)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srav_epi32(__m512i __W, __mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_srav_epi32(__X, __Y),
-                                           (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srav_epi32(__mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_srav_epi32(__X, __Y),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srav_epi64(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psrav8di((__v8di)__X, (__v8di)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srav_epi64(__m512i __W, __mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_srav_epi64(__X, __Y),
-                                            (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srav_epi64(__mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_srav_epi64(__X, __Y),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srl_epi32(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_psrld512((__v16si) __A, (__v4si)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srl_epi32(__m512i __W, __mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_srl_epi32(__A, __B),
-                                          (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srl_epi32(__mmask16 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                          (__v16si)_mm512_srl_epi32(__A, __B),
-                                          (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srl_epi64(__m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_psrlq512((__v8di)__A, (__v2di)__B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srl_epi64(__m512i __W, __mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                           (__v8di)_mm512_srl_epi64(__A, __B),
-                                           (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srl_epi64(__mmask8 __U, __m512i __A, __m128i __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                           (__v8di)_mm512_srl_epi64(__A, __B),
-                                           (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srlv_epi32(__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psrlv16si((__v16si)__X, (__v16si)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srlv_epi32(__m512i __W, __mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_srlv_epi32(__X, __Y),
-                                           (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srlv_epi32(__mmask16 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                           (__v16si)_mm512_srlv_epi32(__X, __Y),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srlv_epi64 (__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_psrlv8di((__v8di)__X, (__v8di)__Y);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srlv_epi64(__m512i __W, __mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_srlv_epi64(__X, __Y),
-                                            (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srlv_epi64(__mmask8 __U, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                            (__v8di)_mm512_srlv_epi64(__X, __Y),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-#define _mm512_ternarylogic_epi32(A, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogd512_mask((__v16si)(__m512i)(A), \
-                                            (__v16si)(__m512i)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)-1)
-
-#define _mm512_mask_ternarylogic_epi32(A, U, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogd512_mask((__v16si)(__m512i)(A), \
-                                            (__v16si)(__m512i)(B), \
-                                            (__v16si)(__m512i)(C), (int)(imm), \
-                                            (__mmask16)(U))
-
-#define _mm512_maskz_ternarylogic_epi32(U, A, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogd512_maskz((__v16si)(__m512i)(A), \
-                                             (__v16si)(__m512i)(B), \
-                                             (__v16si)(__m512i)(C), \
-                                             (int)(imm), (__mmask16)(U))
-
-#define _mm512_ternarylogic_epi64(A, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogq512_mask((__v8di)(__m512i)(A), \
-                                            (__v8di)(__m512i)(B), \
-                                            (__v8di)(__m512i)(C), (int)(imm), \
-                                            (__mmask8)-1)
-
-#define _mm512_mask_ternarylogic_epi64(A, U, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogq512_mask((__v8di)(__m512i)(A), \
-                                            (__v8di)(__m512i)(B), \
-                                            (__v8di)(__m512i)(C), (int)(imm), \
-                                            (__mmask8)(U))
-
-#define _mm512_maskz_ternarylogic_epi64(U, A, B, C, imm) \
-  (__m512i)__builtin_ia32_pternlogq512_maskz((__v8di)(__m512i)(A), \
-                                             (__v8di)(__m512i)(B), \
-                                             (__v8di)(__m512i)(C), (int)(imm), \
-                                             (__mmask8)(U))
-
-#ifdef __x86_64__
-#define _mm_cvt_roundsd_i64(A, R) \
-  (long long)__builtin_ia32_vcvtsd2si64((__v2df)(__m128d)(A), (int)(R))
-#endif
-
-#define _mm_cvt_roundsd_si32(A, R) \
-  (int)__builtin_ia32_vcvtsd2si32((__v2df)(__m128d)(A), (int)(R))
-
-#define _mm_cvt_roundsd_i32(A, R) \
-  (int)__builtin_ia32_vcvtsd2si32((__v2df)(__m128d)(A), (int)(R))
-
-#define _mm_cvt_roundsd_u32(A, R) \
-  (unsigned int)__builtin_ia32_vcvtsd2usi32((__v2df)(__m128d)(A), (int)(R))
-
-static __inline__ unsigned __DEFAULT_FN_ATTRS128
-_mm_cvtsd_u32 (__m128d __A)
-{
-  return (unsigned) __builtin_ia32_vcvtsd2usi32 ((__v2df) __A,
-             _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvt_roundsd_u64(A, R) \
-  (unsigned long long)__builtin_ia32_vcvtsd2usi64((__v2df)(__m128d)(A), \
-                                                  (int)(R))
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS128
-_mm_cvtsd_u64 (__m128d __A)
-{
-  return (unsigned long long) __builtin_ia32_vcvtsd2usi64 ((__v2df)
-                 __A,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm_cvt_roundss_si32(A, R) \
-  (int)__builtin_ia32_vcvtss2si32((__v4sf)(__m128)(A), (int)(R))
-
-#define _mm_cvt_roundss_i32(A, R) \
-  (int)__builtin_ia32_vcvtss2si32((__v4sf)(__m128)(A), (int)(R))
-
-#ifdef __x86_64__
-#define _mm_cvt_roundss_si64(A, R) \
-  (long long)__builtin_ia32_vcvtss2si64((__v4sf)(__m128)(A), (int)(R))
-
-#define _mm_cvt_roundss_i64(A, R) \
-  (long long)__builtin_ia32_vcvtss2si64((__v4sf)(__m128)(A), (int)(R))
-#endif
-
-#define _mm_cvt_roundss_u32(A, R) \
-  (unsigned int)__builtin_ia32_vcvtss2usi32((__v4sf)(__m128)(A), (int)(R))
-
-static __inline__ unsigned __DEFAULT_FN_ATTRS128
-_mm_cvtss_u32 (__m128 __A)
-{
-  return (unsigned) __builtin_ia32_vcvtss2usi32 ((__v4sf) __A,
-             _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvt_roundss_u64(A, R) \
-  (unsigned long long)__builtin_ia32_vcvtss2usi64((__v4sf)(__m128)(A), \
-                                                  (int)(R))
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS128
-_mm_cvtss_u64 (__m128 __A)
-{
-  return (unsigned long long) __builtin_ia32_vcvtss2usi64 ((__v4sf)
-                 __A,
-                 _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm_cvtt_roundsd_i32(A, R) \
-  (int)__builtin_ia32_vcvttsd2si32((__v2df)(__m128d)(A), (int)(R))
-
-#define _mm_cvtt_roundsd_si32(A, R) \
-  (int)__builtin_ia32_vcvttsd2si32((__v2df)(__m128d)(A), (int)(R))
-
-static __inline__ int __DEFAULT_FN_ATTRS128
-_mm_cvttsd_i32 (__m128d __A)
-{
-  return (int) __builtin_ia32_vcvttsd2si32 ((__v2df) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvtt_roundsd_si64(A, R) \
-  (long long)__builtin_ia32_vcvttsd2si64((__v2df)(__m128d)(A), (int)(R))
-
-#define _mm_cvtt_roundsd_i64(A, R) \
-  (long long)__builtin_ia32_vcvttsd2si64((__v2df)(__m128d)(A), (int)(R))
-
-static __inline__ long long __DEFAULT_FN_ATTRS128
-_mm_cvttsd_i64 (__m128d __A)
-{
-  return (long long) __builtin_ia32_vcvttsd2si64 ((__v2df) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm_cvtt_roundsd_u32(A, R) \
-  (unsigned int)__builtin_ia32_vcvttsd2usi32((__v2df)(__m128d)(A), (int)(R))
-
-static __inline__ unsigned __DEFAULT_FN_ATTRS128
-_mm_cvttsd_u32 (__m128d __A)
-{
-  return (unsigned) __builtin_ia32_vcvttsd2usi32 ((__v2df) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvtt_roundsd_u64(A, R) \
-  (unsigned long long)__builtin_ia32_vcvttsd2usi64((__v2df)(__m128d)(A), \
-                                                   (int)(R))
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS128
-_mm_cvttsd_u64 (__m128d __A)
-{
-  return (unsigned long long) __builtin_ia32_vcvttsd2usi64 ((__v2df)
-                  __A,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm_cvtt_roundss_i32(A, R) \
-  (int)__builtin_ia32_vcvttss2si32((__v4sf)(__m128)(A), (int)(R))
-
-#define _mm_cvtt_roundss_si32(A, R) \
-  (int)__builtin_ia32_vcvttss2si32((__v4sf)(__m128)(A), (int)(R))
-
-static __inline__ int __DEFAULT_FN_ATTRS128
-_mm_cvttss_i32 (__m128 __A)
-{
-  return (int) __builtin_ia32_vcvttss2si32 ((__v4sf) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvtt_roundss_i64(A, R) \
-  (long long)__builtin_ia32_vcvttss2si64((__v4sf)(__m128)(A), (int)(R))
-
-#define _mm_cvtt_roundss_si64(A, R) \
-  (long long)__builtin_ia32_vcvttss2si64((__v4sf)(__m128)(A), (int)(R))
-
-static __inline__ long long __DEFAULT_FN_ATTRS128
-_mm_cvttss_i64 (__m128 __A)
-{
-  return (long long) __builtin_ia32_vcvttss2si64 ((__v4sf) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm_cvtt_roundss_u32(A, R) \
-  (unsigned int)__builtin_ia32_vcvttss2usi32((__v4sf)(__m128)(A), (int)(R))
-
-static __inline__ unsigned __DEFAULT_FN_ATTRS128
-_mm_cvttss_u32 (__m128 __A)
-{
-  return (unsigned) __builtin_ia32_vcvttss2usi32 ((__v4sf) __A,
-              _MM_FROUND_CUR_DIRECTION);
-}
-
-#ifdef __x86_64__
-#define _mm_cvtt_roundss_u64(A, R) \
-  (unsigned long long)__builtin_ia32_vcvttss2usi64((__v4sf)(__m128)(A), \
-                                                   (int)(R))
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS128
-_mm_cvttss_u64 (__m128 __A)
-{
-  return (unsigned long long) __builtin_ia32_vcvttss2usi64 ((__v4sf)
-                  __A,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-#endif
-
-#define _mm512_permute_pd(X, C) \
-  (__m512d)__builtin_ia32_vpermilpd512((__v8df)(__m512d)(X), (int)(C))
-
-#define _mm512_mask_permute_pd(W, U, X, C) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_permute_pd((X), (C)), \
-                                       (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_permute_pd(U, X, C) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_permute_pd((X), (C)), \
-                                       (__v8df)_mm512_setzero_pd())
-
-#define _mm512_permute_ps(X, C) \
-  (__m512)__builtin_ia32_vpermilps512((__v16sf)(__m512)(X), (int)(C))
-
-#define _mm512_mask_permute_ps(W, U, X, C) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_permute_ps((X), (C)), \
-                                      (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_permute_ps(U, X, C) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_permute_ps((X), (C)), \
-                                      (__v16sf)_mm512_setzero_ps())
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_permutevar_pd(__m512d __A, __m512i __C)
-{
-  return (__m512d)__builtin_ia32_vpermilvarpd512((__v8df)__A, (__v8di)__C);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_permutevar_pd(__m512d __W, __mmask8 __U, __m512d __A, __m512i __C)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                         (__v8df)_mm512_permutevar_pd(__A, __C),
-                                         (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutevar_pd(__mmask8 __U, __m512d __A, __m512i __C)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                         (__v8df)_mm512_permutevar_pd(__A, __C),
-                                         (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_permutevar_ps(__m512 __A, __m512i __C)
-{
-  return (__m512)__builtin_ia32_vpermilvarps512((__v16sf)__A, (__v16si)__C);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_permutevar_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512i __C)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                        (__v16sf)_mm512_permutevar_ps(__A, __C),
-                                        (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutevar_ps(__mmask16 __U, __m512 __A, __m512i __C)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                        (__v16sf)_mm512_permutevar_ps(__A, __C),
-                                        (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline __m512d __DEFAULT_FN_ATTRS512
-_mm512_permutex2var_pd(__m512d __A, __m512i __I, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_vpermi2varpd512((__v8df)__A, (__v8di)__I,
-                                                 (__v8df)__B);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_permutex2var_pd(__m512d __A, __mmask8 __U, __m512i __I, __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                  (__v8df)_mm512_permutex2var_pd(__A, __I, __B),
-                                  (__v8df)__A);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask2_permutex2var_pd(__m512d __A, __m512i __I, __mmask8 __U,
-                             __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                  (__v8df)_mm512_permutex2var_pd(__A, __I, __B),
-                                  (__v8df)(__m512d)__I);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutex2var_pd(__mmask8 __U, __m512d __A, __m512i __I,
-                             __m512d __B)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__U,
-                                  (__v8df)_mm512_permutex2var_pd(__A, __I, __B),
-                                  (__v8df)_mm512_setzero_pd());
-}
-
-static __inline __m512 __DEFAULT_FN_ATTRS512
-_mm512_permutex2var_ps(__m512 __A, __m512i __I, __m512 __B)
-{
-  return (__m512)__builtin_ia32_vpermi2varps512((__v16sf)__A, (__v16si)__I,
-                                                (__v16sf) __B);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_permutex2var_ps(__m512 __A, __mmask16 __U, __m512i __I, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                 (__v16sf)_mm512_permutex2var_ps(__A, __I, __B),
-                                 (__v16sf)__A);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask2_permutex2var_ps(__m512 __A, __m512i __I, __mmask16 __U, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                 (__v16sf)_mm512_permutex2var_ps(__A, __I, __B),
-                                 (__v16sf)(__m512)__I);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutex2var_ps(__mmask16 __U, __m512 __A, __m512i __I, __m512 __B)
-{
-  return (__m512)__builtin_ia32_selectps_512(__U,
-                                 (__v16sf)_mm512_permutex2var_ps(__A, __I, __B),
-                                 (__v16sf)_mm512_setzero_ps());
-}
-
-
-#define _mm512_cvtt_roundpd_epu32(A, R) \
-  (__m256i)__builtin_ia32_cvttpd2udq512_mask((__v8df)(__m512d)(A), \
-                                             (__v8si)_mm256_undefined_si256(), \
-                                             (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvtt_roundpd_epu32(W, U, A, R) \
-  (__m256i)__builtin_ia32_cvttpd2udq512_mask((__v8df)(__m512d)(A), \
-                                             (__v8si)(__m256i)(W), \
-                                             (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_cvtt_roundpd_epu32(U, A, R) \
-  (__m256i)__builtin_ia32_cvttpd2udq512_mask((__v8df)(__m512d)(A), \
-                                             (__v8si)_mm256_setzero_si256(), \
-                                             (__mmask8)(U), (int)(R))
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvttpd_epu32 (__m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvttpd2udq512_mask ((__v8df) __A,
-                  (__v8si)
-                  _mm256_undefined_si256 (),
-                  (__mmask8) -1,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvttpd_epu32 (__m256i __W, __mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvttpd2udq512_mask ((__v8df) __A,
-                  (__v8si) __W,
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvttpd_epu32 (__mmask8 __U, __m512d __A)
-{
-  return (__m256i) __builtin_ia32_cvttpd2udq512_mask ((__v8df) __A,
-                  (__v8si)
-                  _mm256_setzero_si256 (),
-                  (__mmask8) __U,
-                  _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_roundscale_round_sd(A, B, imm, R) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)_mm_setzero_pd(), \
-                                                (__mmask8)-1, (int)(imm), \
-                                                (int)(R))
-
-#define _mm_roundscale_sd(A, B, imm) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)_mm_setzero_pd(), \
-                                                (__mmask8)-1, (int)(imm), \
-                                                _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_roundscale_sd(W, U, A, B, imm) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)(__m128d)(W), \
-                                                (__mmask8)(U), (int)(imm), \
-                                                _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_roundscale_round_sd(W, U, A, B, I, R) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)(__m128d)(W), \
-                                                (__mmask8)(U), (int)(I), \
-                                                (int)(R))
-
-#define _mm_maskz_roundscale_sd(U, A, B, I) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)_mm_setzero_pd(), \
-                                                (__mmask8)(U), (int)(I), \
-                                                _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_roundscale_round_sd(U, A, B, I, R) \
-  (__m128d)__builtin_ia32_rndscalesd_round_mask((__v2df)(__m128d)(A), \
-                                                (__v2df)(__m128d)(B), \
-                                                (__v2df)_mm_setzero_pd(), \
-                                                (__mmask8)(U), (int)(I), \
-                                                (int)(R))
-
-#define _mm_roundscale_round_ss(A, B, imm, R) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)_mm_setzero_ps(), \
-                                               (__mmask8)-1, (int)(imm), \
-                                               (int)(R))
-
-#define _mm_roundscale_ss(A, B, imm) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)_mm_setzero_ps(), \
-                                               (__mmask8)-1, (int)(imm), \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_roundscale_ss(W, U, A, B, I) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)(__m128)(W), \
-                                               (__mmask8)(U), (int)(I), \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_roundscale_round_ss(W, U, A, B, I, R) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)(__m128)(W), \
-                                               (__mmask8)(U), (int)(I), \
-                                               (int)(R))
-
-#define _mm_maskz_roundscale_ss(U, A, B, I) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)_mm_setzero_ps(), \
-                                               (__mmask8)(U), (int)(I), \
-                                               _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_maskz_roundscale_round_ss(U, A, B, I, R) \
-  (__m128)__builtin_ia32_rndscaless_round_mask((__v4sf)(__m128)(A), \
-                                               (__v4sf)(__m128)(B), \
-                                               (__v4sf)_mm_setzero_ps(), \
-                                               (__mmask8)(U), (int)(I), \
-                                               (int)(R))
-
-#define _mm512_scalef_round_pd(A, B, R) \
-  (__m512d)__builtin_ia32_scalefpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)_mm512_undefined_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_scalef_round_pd(W, U, A, B, R) \
-  (__m512d)__builtin_ia32_scalefpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)(__m512d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_scalef_round_pd(U, A, B, R) \
-  (__m512d)__builtin_ia32_scalefpd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(B), \
-                                           (__v8df)_mm512_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_scalef_pd (__m512d __A, __m512d __B)
-{
-  return (__m512d) __builtin_ia32_scalefpd512_mask ((__v8df) __A,
-                (__v8df) __B,
-                (__v8df)
-                _mm512_undefined_pd (),
-                (__mmask8) -1,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_scalef_pd (__m512d __W, __mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d) __builtin_ia32_scalefpd512_mask ((__v8df) __A,
-                (__v8df) __B,
-                (__v8df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_scalef_pd (__mmask8 __U, __m512d __A, __m512d __B)
-{
-  return (__m512d) __builtin_ia32_scalefpd512_mask ((__v8df) __A,
-                (__v8df) __B,
-                (__v8df)
-                _mm512_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_scalef_round_ps(A, B, R) \
-  (__m512)__builtin_ia32_scalefps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          (__v16sf)_mm512_undefined_ps(), \
-                                          (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_scalef_round_ps(W, U, A, B, R) \
-  (__m512)__builtin_ia32_scalefps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          (__v16sf)(__m512)(W), \
-                                          (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_scalef_round_ps(U, A, B, R) \
-  (__m512)__builtin_ia32_scalefps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(B), \
-                                          (__v16sf)_mm512_setzero_ps(), \
-                                          (__mmask16)(U), (int)(R))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_scalef_ps (__m512 __A, __m512 __B)
-{
-  return (__m512) __builtin_ia32_scalefps512_mask ((__v16sf) __A,
-               (__v16sf) __B,
-               (__v16sf)
-               _mm512_undefined_ps (),
-               (__mmask16) -1,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_scalef_ps (__m512 __W, __mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512) __builtin_ia32_scalefps512_mask ((__v16sf) __A,
-               (__v16sf) __B,
-               (__v16sf) __W,
-               (__mmask16) __U,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_scalef_ps (__mmask16 __U, __m512 __A, __m512 __B)
-{
-  return (__m512) __builtin_ia32_scalefps512_mask ((__v16sf) __A,
-               (__v16sf) __B,
-               (__v16sf)
-               _mm512_setzero_ps (),
-               (__mmask16) __U,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_scalef_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_scalefsd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v2df)(__m128d)(B), \
-                                              (__v2df)_mm_setzero_pd(), \
-                                              (__mmask8)-1, (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_scalef_sd (__m128d __A, __m128d __B)
-{
-  return (__m128d) __builtin_ia32_scalefsd_round_mask ((__v2df) __A,
-              (__v2df)( __B), (__v2df) _mm_setzero_pd(),
-              (__mmask8) -1,
-              _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_scalef_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_scalefsd_round_mask ( (__v2df) __A,
-                 (__v2df) __B,
-                (__v2df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_scalef_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_scalefsd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v2df)(__m128d)(B), \
-                                              (__v2df)(__m128d)(W), \
-                                              (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_scalef_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_scalefsd_round_mask ( (__v2df) __A,
-                 (__v2df) __B,
-                (__v2df) _mm_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_scalef_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_scalefsd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v2df)(__m128d)(B), \
-                                              (__v2df)_mm_setzero_pd(), \
-                                              (__mmask8)(U), (int)(R))
-
-#define _mm_scalef_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_scalefss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v4sf)(__m128)(B), \
-                                             (__v4sf)_mm_setzero_ps(), \
-                                             (__mmask8)-1, (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_scalef_ss (__m128 __A, __m128 __B)
-{
-  return (__m128) __builtin_ia32_scalefss_round_mask ((__v4sf) __A,
-             (__v4sf)( __B), (__v4sf) _mm_setzero_ps(),
-             (__mmask8) -1,
-             _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_scalef_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_scalefss_round_mask ( (__v4sf) __A,
-                (__v4sf) __B,
-                (__v4sf) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_scalef_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_scalefss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v4sf)(__m128)(B), \
-                                             (__v4sf)(__m128)(W), \
-                                             (__mmask8)(U), (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_scalef_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_scalefss_round_mask ( (__v4sf) __A,
-                 (__v4sf) __B,
-                (__v4sf) _mm_setzero_ps (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_scalef_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_scalefss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v4sf)(__m128)(B), \
-                                             (__v4sf)_mm_setzero_ps(), \
-                                             (__mmask8)(U), \
-                                             (int)(R))
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srai_epi32(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_psradi512((__v16si)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srai_epi32(__m512i __W, __mmask16 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_srai_epi32(__A, __B),
-                                         (__v16si)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srai_epi32(__mmask16 __U, __m512i __A, int __B) {
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U,
-                                         (__v16si)_mm512_srai_epi32(__A, __B),
-                                         (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_srai_epi64(__m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_psraqi512((__v8di)__A, __B);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_srai_epi64(__m512i __W, __mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_srai_epi64(__A, __B),
-                                          (__v8di)__W);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_srai_epi64(__mmask8 __U, __m512i __A, int __B)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U,
-                                          (__v8di)_mm512_srai_epi64(__A, __B),
-                                          (__v8di)_mm512_setzero_si512());
-}
-
-#define _mm512_shuffle_f32x4(A, B, imm) \
-  (__m512)__builtin_ia32_shuf_f32x4((__v16sf)(__m512)(A), \
-                                    (__v16sf)(__m512)(B), (int)(imm))
-
-#define _mm512_mask_shuffle_f32x4(W, U, A, B, imm) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_shuffle_f32x4((A), (B), (imm)), \
-                                      (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_shuffle_f32x4(U, A, B, imm) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_shuffle_f32x4((A), (B), (imm)), \
-                                      (__v16sf)_mm512_setzero_ps())
-
-#define _mm512_shuffle_f64x2(A, B, imm) \
-  (__m512d)__builtin_ia32_shuf_f64x2((__v8df)(__m512d)(A), \
-                                     (__v8df)(__m512d)(B), (int)(imm))
-
-#define _mm512_mask_shuffle_f64x2(W, U, A, B, imm) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_shuffle_f64x2((A), (B), (imm)), \
-                                       (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_shuffle_f64x2(U, A, B, imm) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_shuffle_f64x2((A), (B), (imm)), \
-                                       (__v8df)_mm512_setzero_pd())
-
-#define _mm512_shuffle_i32x4(A, B, imm) \
-  (__m512i)__builtin_ia32_shuf_i32x4((__v16si)(__m512i)(A), \
-                                     (__v16si)(__m512i)(B), (int)(imm))
-
-#define _mm512_mask_shuffle_i32x4(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_shuffle_i32x4((A), (B), (imm)), \
-                                      (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_shuffle_i32x4(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_shuffle_i32x4((A), (B), (imm)), \
-                                      (__v16si)_mm512_setzero_si512())
-
-#define _mm512_shuffle_i64x2(A, B, imm) \
-  (__m512i)__builtin_ia32_shuf_i64x2((__v8di)(__m512i)(A), \
-                                     (__v8di)(__m512i)(B), (int)(imm))
-
-#define _mm512_mask_shuffle_i64x2(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_shuffle_i64x2((A), (B), (imm)), \
-                                      (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_shuffle_i64x2(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_shuffle_i64x2((A), (B), (imm)), \
-                                      (__v8di)_mm512_setzero_si512())
-
-#define _mm512_shuffle_pd(A, B, M) \
-  (__m512d)__builtin_ia32_shufpd512((__v8df)(__m512d)(A), \
-                                    (__v8df)(__m512d)(B), (int)(M))
-
-#define _mm512_mask_shuffle_pd(W, U, A, B, M) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_shuffle_pd((A), (B), (M)), \
-                                       (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_shuffle_pd(U, A, B, M) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_shuffle_pd((A), (B), (M)), \
-                                       (__v8df)_mm512_setzero_pd())
-
-#define _mm512_shuffle_ps(A, B, M) \
-  (__m512)__builtin_ia32_shufps512((__v16sf)(__m512)(A), \
-                                   (__v16sf)(__m512)(B), (int)(M))
-
-#define _mm512_mask_shuffle_ps(W, U, A, B, M) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_shuffle_ps((A), (B), (M)), \
-                                      (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_shuffle_ps(U, A, B, M) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                      (__v16sf)_mm512_shuffle_ps((A), (B), (M)), \
-                                      (__v16sf)_mm512_setzero_ps())
-
-#define _mm_sqrt_round_sd(A, B, R) \
-  (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
-                                            (__v2df)(__m128d)(B), \
-                                            (__v2df)_mm_setzero_pd(), \
-                                            (__mmask8)-1, (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_sqrt_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __A,
-                 (__v2df) __B,
-                (__v2df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_sqrt_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
-                                            (__v2df)(__m128d)(B), \
-                                            (__v2df)(__m128d)(W), \
-                                            (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_sqrt_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
- return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __A,
-                 (__v2df) __B,
-                (__v2df) _mm_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_sqrt_round_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
-                                            (__v2df)(__m128d)(B), \
-                                            (__v2df)_mm_setzero_pd(), \
-                                            (__mmask8)(U), (int)(R))
-
-#define _mm_sqrt_round_ss(A, B, R) \
-  (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
-                                           (__v4sf)(__m128)(B), \
-                                           (__v4sf)_mm_setzero_ps(), \
-                                           (__mmask8)-1, (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_sqrt_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __A,
-                 (__v4sf) __B,
-                (__v4sf) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask_sqrt_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
-                                           (__v4sf)(__m128)(B), \
-                                           (__v4sf)(__m128)(W), (__mmask8)(U), \
-                                           (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_sqrt_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
- return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __A,
-                 (__v4sf) __B,
-                (__v4sf) _mm_setzero_ps (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_sqrt_round_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
-                                           (__v4sf)(__m128)(B), \
-                                           (__v4sf)_mm_setzero_ps(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_broadcast_f32x4(__m128 __A)
-{
-  return (__m512)__builtin_shufflevector((__v4sf)__A, (__v4sf)__A,
-                                         0, 1, 2, 3, 0, 1, 2, 3,
-                                         0, 1, 2, 3, 0, 1, 2, 3);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcast_f32x4(__m512 __O, __mmask16 __M, __m128 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__M,
-                                           (__v16sf)_mm512_broadcast_f32x4(__A),
-                                           (__v16sf)__O);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcast_f32x4(__mmask16 __M, __m128 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__M,
-                                           (__v16sf)_mm512_broadcast_f32x4(__A),
-                                           (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_broadcast_f64x4(__m256d __A)
-{
-  return (__m512d)__builtin_shufflevector((__v4df)__A, (__v4df)__A,
-                                          0, 1, 2, 3, 0, 1, 2, 3);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcast_f64x4(__m512d __O, __mmask8 __M, __m256d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__M,
-                                            (__v8df)_mm512_broadcast_f64x4(__A),
-                                            (__v8df)__O);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcast_f64x4(__mmask8 __M, __m256d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__M,
-                                            (__v8df)_mm512_broadcast_f64x4(__A),
-                                            (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_broadcast_i32x4(__m128i __A)
-{
-  return (__m512i)__builtin_shufflevector((__v4si)__A, (__v4si)__A,
-                                          0, 1, 2, 3, 0, 1, 2, 3,
-                                          0, 1, 2, 3, 0, 1, 2, 3);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcast_i32x4(__m512i __O, __mmask16 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                           (__v16si)_mm512_broadcast_i32x4(__A),
-                                           (__v16si)__O);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcast_i32x4(__mmask16 __M, __m128i __A)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                           (__v16si)_mm512_broadcast_i32x4(__A),
-                                           (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_broadcast_i64x4(__m256i __A)
-{
-  return (__m512i)__builtin_shufflevector((__v4di)__A, (__v4di)__A,
-                                          0, 1, 2, 3, 0, 1, 2, 3);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcast_i64x4(__m512i __O, __mmask8 __M, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                            (__v8di)_mm512_broadcast_i64x4(__A),
-                                            (__v8di)__O);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcast_i64x4(__mmask8 __M, __m256i __A)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                            (__v8di)_mm512_broadcast_i64x4(__A),
-                                            (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcastsd_pd (__m512d __O, __mmask8 __M, __m128d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__M,
-                                              (__v8df) _mm512_broadcastsd_pd(__A),
-                                              (__v8df) __O);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcastsd_pd (__mmask8 __M, __m128d __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512(__M,
-                                              (__v8df) _mm512_broadcastsd_pd(__A),
-                                              (__v8df) _mm512_setzero_pd());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_broadcastss_ps (__m512 __O, __mmask16 __M, __m128 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512(__M,
-                                             (__v16sf) _mm512_broadcastss_ps(__A),
-                                             (__v16sf) __O);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_broadcastss_ps (__mmask16 __M, __m128 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512(__M,
-                                             (__v16sf) _mm512_broadcastss_ps(__A),
-                                             (__v16sf) _mm512_setzero_ps());
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtsepi32_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsdb512_mask ((__v16si) __A,
-               (__v16qi) _mm_undefined_si128 (),
-               (__mmask16) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi32_epi8 (__m128i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsdb512_mask ((__v16si) __A,
-               (__v16qi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtsepi32_epi8 (__mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsdb512_mask ((__v16si) __A,
-               (__v16qi) _mm_setzero_si128 (),
-               __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi32_storeu_epi8 (void * __P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovsdb512mem_mask ((__v16qi *) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtsepi32_epi16 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsdw512_mask ((__v16si) __A,
-               (__v16hi) _mm256_undefined_si256 (),
-               (__mmask16) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi32_epi16 (__m256i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsdw512_mask ((__v16si) __A,
-               (__v16hi) __O, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtsepi32_epi16 (__mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsdw512_mask ((__v16si) __A,
-               (__v16hi) _mm256_setzero_si256 (),
-               __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi32_storeu_epi16 (void *__P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovsdw512mem_mask ((__v16hi*) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtsepi64_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqb512_mask ((__v8di) __A,
-               (__v16qi) _mm_undefined_si128 (),
-               (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_epi8 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqb512_mask ((__v8di) __A,
-               (__v16qi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtsepi64_epi8 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqb512_mask ((__v8di) __A,
-               (__v16qi) _mm_setzero_si128 (),
-               __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_storeu_epi8 (void * __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovsqb512mem_mask ((__v16qi *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtsepi64_epi32 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsqd512_mask ((__v8di) __A,
-               (__v8si) _mm256_undefined_si256 (),
-               (__mmask8) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_epi32 (__m256i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsqd512_mask ((__v8di) __A,
-               (__v8si) __O, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtsepi64_epi32 (__mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovsqd512_mask ((__v8di) __A,
-               (__v8si) _mm256_setzero_si256 (),
-               __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_storeu_epi32 (void *__P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovsqd512mem_mask ((__v8si *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtsepi64_epi16 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqw512_mask ((__v8di) __A,
-               (__v8hi) _mm_undefined_si128 (),
-               (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_epi16 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqw512_mask ((__v8di) __A,
-               (__v8hi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtsepi64_epi16 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovsqw512_mask ((__v8di) __A,
-               (__v8hi) _mm_setzero_si128 (),
-               __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtsepi64_storeu_epi16 (void * __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovsqw512mem_mask ((__v8hi *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtusepi32_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusdb512_mask ((__v16si) __A,
-                (__v16qi) _mm_undefined_si128 (),
-                (__mmask16) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi32_epi8 (__m128i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusdb512_mask ((__v16si) __A,
-                (__v16qi) __O,
-                __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtusepi32_epi8 (__mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusdb512_mask ((__v16si) __A,
-                (__v16qi) _mm_setzero_si128 (),
-                __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi32_storeu_epi8 (void * __P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovusdb512mem_mask ((__v16qi *) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtusepi32_epi16 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusdw512_mask ((__v16si) __A,
-                (__v16hi) _mm256_undefined_si256 (),
-                (__mmask16) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi32_epi16 (__m256i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusdw512_mask ((__v16si) __A,
-                (__v16hi) __O,
-                __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtusepi32_epi16 (__mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusdw512_mask ((__v16si) __A,
-                (__v16hi) _mm256_setzero_si256 (),
-                __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi32_storeu_epi16 (void *__P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovusdw512mem_mask ((__v16hi*) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtusepi64_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqb512_mask ((__v8di) __A,
-                (__v16qi) _mm_undefined_si128 (),
-                (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_epi8 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqb512_mask ((__v8di) __A,
-                (__v16qi) __O,
-                __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtusepi64_epi8 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqb512_mask ((__v8di) __A,
-                (__v16qi) _mm_setzero_si128 (),
-                __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_storeu_epi8 (void * __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovusqb512mem_mask ((__v16qi *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtusepi64_epi32 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusqd512_mask ((__v8di) __A,
-                (__v8si) _mm256_undefined_si256 (),
-                (__mmask8) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_epi32 (__m256i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusqd512_mask ((__v8di) __A,
-                (__v8si) __O, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtusepi64_epi32 (__mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovusqd512_mask ((__v8di) __A,
-                (__v8si) _mm256_setzero_si256 (),
-                __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_storeu_epi32 (void* __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovusqd512mem_mask ((__v8si*) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtusepi64_epi16 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqw512_mask ((__v8di) __A,
-                (__v8hi) _mm_undefined_si128 (),
-                (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_epi16 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqw512_mask ((__v8di) __A,
-                (__v8hi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtusepi64_epi16 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovusqw512_mask ((__v8di) __A,
-                (__v8hi) _mm_setzero_si128 (),
-                __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtusepi64_storeu_epi16 (void *__P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovusqw512mem_mask ((__v8hi*) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovdb512_mask ((__v16si) __A,
-              (__v16qi) _mm_undefined_si128 (),
-              (__mmask16) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_epi8 (__m128i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovdb512_mask ((__v16si) __A,
-              (__v16qi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi32_epi8 (__mmask16 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovdb512_mask ((__v16si) __A,
-              (__v16qi) _mm_setzero_si128 (),
-              __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_storeu_epi8 (void * __P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovdb512mem_mask ((__v16qi *) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi32_epi16 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovdw512_mask ((__v16si) __A,
-              (__v16hi) _mm256_undefined_si256 (),
-              (__mmask16) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_epi16 (__m256i __O, __mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovdw512_mask ((__v16si) __A,
-              (__v16hi) __O, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi32_epi16 (__mmask16 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovdw512_mask ((__v16si) __A,
-              (__v16hi) _mm256_setzero_si256 (),
-              __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi32_storeu_epi16 (void * __P, __mmask16 __M, __m512i __A)
-{
-  __builtin_ia32_pmovdw512mem_mask ((__v16hi *) __P, (__v16si) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi64_epi8 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqb512_mask ((__v8di) __A,
-              (__v16qi) _mm_undefined_si128 (),
-              (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_epi8 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqb512_mask ((__v8di) __A,
-              (__v16qi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi64_epi8 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqb512_mask ((__v8di) __A,
-              (__v16qi) _mm_setzero_si128 (),
-              __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_storeu_epi8 (void * __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovqb512mem_mask ((__v16qi *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi64_epi32 (__m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovqd512_mask ((__v8di) __A,
-              (__v8si) _mm256_undefined_si256 (),
-              (__mmask8) -1);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_epi32 (__m256i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovqd512_mask ((__v8di) __A,
-              (__v8si) __O, __M);
-}
-
-static __inline__ __m256i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi64_epi32 (__mmask8 __M, __m512i __A)
-{
-  return (__m256i) __builtin_ia32_pmovqd512_mask ((__v8di) __A,
-              (__v8si) _mm256_setzero_si256 (),
-              __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_storeu_epi32 (void* __P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovqd512mem_mask ((__v8si *) __P, (__v8di) __A, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_cvtepi64_epi16 (__m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqw512_mask ((__v8di) __A,
-              (__v8hi) _mm_undefined_si128 (),
-              (__mmask8) -1);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_epi16 (__m128i __O, __mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqw512_mask ((__v8di) __A,
-              (__v8hi) __O, __M);
-}
-
-static __inline__ __m128i __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtepi64_epi16 (__mmask8 __M, __m512i __A)
-{
-  return (__m128i) __builtin_ia32_pmovqw512_mask ((__v8di) __A,
-              (__v8hi) _mm_setzero_si128 (),
-              __M);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtepi64_storeu_epi16 (void *__P, __mmask8 __M, __m512i __A)
-{
-  __builtin_ia32_pmovqw512mem_mask ((__v8hi *) __P, (__v8di) __A, __M);
-}
-
-#define _mm512_extracti32x4_epi32(A, imm) \
-  (__m128i)__builtin_ia32_extracti32x4_mask((__v16si)(__m512i)(A), (int)(imm), \
-                                            (__v4si)_mm_undefined_si128(), \
-                                            (__mmask8)-1)
-
-#define _mm512_mask_extracti32x4_epi32(W, U, A, imm) \
-  (__m128i)__builtin_ia32_extracti32x4_mask((__v16si)(__m512i)(A), (int)(imm), \
-                                            (__v4si)(__m128i)(W), \
-                                            (__mmask8)(U))
-
-#define _mm512_maskz_extracti32x4_epi32(U, A, imm) \
-  (__m128i)__builtin_ia32_extracti32x4_mask((__v16si)(__m512i)(A), (int)(imm), \
-                                            (__v4si)_mm_setzero_si128(), \
-                                            (__mmask8)(U))
-
-#define _mm512_extracti64x4_epi64(A, imm) \
-  (__m256i)__builtin_ia32_extracti64x4_mask((__v8di)(__m512i)(A), (int)(imm), \
-                                            (__v4di)_mm256_undefined_si256(), \
-                                            (__mmask8)-1)
-
-#define _mm512_mask_extracti64x4_epi64(W, U, A, imm) \
-  (__m256i)__builtin_ia32_extracti64x4_mask((__v8di)(__m512i)(A), (int)(imm), \
-                                            (__v4di)(__m256i)(W), \
-                                            (__mmask8)(U))
-
-#define _mm512_maskz_extracti64x4_epi64(U, A, imm) \
-  (__m256i)__builtin_ia32_extracti64x4_mask((__v8di)(__m512i)(A), (int)(imm), \
-                                            (__v4di)_mm256_setzero_si256(), \
-                                            (__mmask8)(U))
-
-#define _mm512_insertf64x4(A, B, imm) \
-  (__m512d)__builtin_ia32_insertf64x4((__v8df)(__m512d)(A), \
-                                      (__v4df)(__m256d)(B), (int)(imm))
-
-#define _mm512_mask_insertf64x4(W, U, A, B, imm) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                  (__v8df)_mm512_insertf64x4((A), (B), (imm)), \
-                                  (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_insertf64x4(U, A, B, imm) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                  (__v8df)_mm512_insertf64x4((A), (B), (imm)), \
-                                  (__v8df)_mm512_setzero_pd())
-
-#define _mm512_inserti64x4(A, B, imm) \
-  (__m512i)__builtin_ia32_inserti64x4((__v8di)(__m512i)(A), \
-                                      (__v4di)(__m256i)(B), (int)(imm))
-
-#define _mm512_mask_inserti64x4(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                  (__v8di)_mm512_inserti64x4((A), (B), (imm)), \
-                                  (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_inserti64x4(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                  (__v8di)_mm512_inserti64x4((A), (B), (imm)), \
-                                  (__v8di)_mm512_setzero_si512())
-
-#define _mm512_insertf32x4(A, B, imm) \
-  (__m512)__builtin_ia32_insertf32x4((__v16sf)(__m512)(A), \
-                                     (__v4sf)(__m128)(B), (int)(imm))
-
-#define _mm512_mask_insertf32x4(W, U, A, B, imm) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                 (__v16sf)_mm512_insertf32x4((A), (B), (imm)), \
-                                 (__v16sf)(__m512)(W))
-
-#define _mm512_maskz_insertf32x4(U, A, B, imm) \
-  (__m512)__builtin_ia32_selectps_512((__mmask16)(U), \
-                                 (__v16sf)_mm512_insertf32x4((A), (B), (imm)), \
-                                 (__v16sf)_mm512_setzero_ps())
-
-#define _mm512_inserti32x4(A, B, imm) \
-  (__m512i)__builtin_ia32_inserti32x4((__v16si)(__m512i)(A), \
-                                      (__v4si)(__m128i)(B), (int)(imm))
-
-#define _mm512_mask_inserti32x4(W, U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                 (__v16si)_mm512_inserti32x4((A), (B), (imm)), \
-                                 (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_inserti32x4(U, A, B, imm) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                 (__v16si)_mm512_inserti32x4((A), (B), (imm)), \
-                                 (__v16si)_mm512_setzero_si512())
-
-#define _mm512_getmant_round_pd(A, B, C, R) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)_mm512_undefined_pd(), \
-                                            (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_getmant_round_pd(W, U, A, B, C, R) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)(__m512d)(W), \
-                                            (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_getmant_round_pd(U, A, B, C, R) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)_mm512_setzero_pd(), \
-                                            (__mmask8)(U), (int)(R))
-
-#define _mm512_getmant_pd(A, B, C) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)_mm512_setzero_pd(), \
-                                            (__mmask8)-1, \
-                                            _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_getmant_pd(W, U, A, B, C) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)(__m512d)(W), \
-                                            (__mmask8)(U), \
-                                            _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_getmant_pd(U, A, B, C) \
-  (__m512d)__builtin_ia32_getmantpd512_mask((__v8df)(__m512d)(A), \
-                                            (int)(((C)<<2) | (B)), \
-                                            (__v8df)_mm512_setzero_pd(), \
-                                            (__mmask8)(U), \
-                                            _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_getmant_round_ps(A, B, C, R) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2) | (B)), \
-                                           (__v16sf)_mm512_undefined_ps(), \
-                                           (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_getmant_round_ps(W, U, A, B, C, R) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2) | (B)), \
-                                           (__v16sf)(__m512)(W), \
-                                           (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_getmant_round_ps(U, A, B, C, R) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2) | (B)), \
-                                           (__v16sf)_mm512_setzero_ps(), \
-                                           (__mmask16)(U), (int)(R))
-
-#define _mm512_getmant_ps(A, B, C) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2)|(B)), \
-                                           (__v16sf)_mm512_undefined_ps(), \
-                                           (__mmask16)-1, \
-                                           _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_mask_getmant_ps(W, U, A, B, C) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2)|(B)), \
-                                           (__v16sf)(__m512)(W), \
-                                           (__mmask16)(U), \
-                                           _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_maskz_getmant_ps(U, A, B, C) \
-  (__m512)__builtin_ia32_getmantps512_mask((__v16sf)(__m512)(A), \
-                                           (int)(((C)<<2)|(B)), \
-                                           (__v16sf)_mm512_setzero_ps(), \
-                                           (__mmask16)(U), \
-                                           _MM_FROUND_CUR_DIRECTION)
-
-#define _mm512_getexp_round_pd(A, R) \
-  (__m512d)__builtin_ia32_getexppd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)_mm512_undefined_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_getexp_round_pd(W, U, A, R) \
-  (__m512d)__builtin_ia32_getexppd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)(__m512d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_getexp_round_pd(U, A, R) \
-  (__m512d)__builtin_ia32_getexppd512_mask((__v8df)(__m512d)(A), \
-                                           (__v8df)_mm512_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_getexp_pd (__m512d __A)
-{
-  return (__m512d) __builtin_ia32_getexppd512_mask ((__v8df) __A,
-                (__v8df) _mm512_undefined_pd (),
-                (__mmask8) -1,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_getexp_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_getexppd512_mask ((__v8df) __A,
-                (__v8df) __W,
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_getexp_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_getexppd512_mask ((__v8df) __A,
-                (__v8df) _mm512_setzero_pd (),
-                (__mmask8) __U,
-                _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_getexp_round_ps(A, R) \
-  (__m512)__builtin_ia32_getexpps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)_mm512_undefined_ps(), \
-                                          (__mmask16)-1, (int)(R))
-
-#define _mm512_mask_getexp_round_ps(W, U, A, R) \
-  (__m512)__builtin_ia32_getexpps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)(__m512)(W), \
-                                          (__mmask16)(U), (int)(R))
-
-#define _mm512_maskz_getexp_round_ps(U, A, R) \
-  (__m512)__builtin_ia32_getexpps512_mask((__v16sf)(__m512)(A), \
-                                          (__v16sf)_mm512_setzero_ps(), \
-                                          (__mmask16)(U), (int)(R))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_getexp_ps (__m512 __A)
-{
-  return (__m512) __builtin_ia32_getexpps512_mask ((__v16sf) __A,
-               (__v16sf) _mm512_undefined_ps (),
-               (__mmask16) -1,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_getexp_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_getexpps512_mask ((__v16sf) __A,
-               (__v16sf) __W,
-               (__mmask16) __U,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_getexp_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_getexpps512_mask ((__v16sf) __A,
-               (__v16sf) _mm512_setzero_ps (),
-               (__mmask16) __U,
-               _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm512_i64gather_ps(index, addr, scale) \
-  (__m256)__builtin_ia32_gatherdiv16sf((__v8sf)_mm256_undefined_ps(), \
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), (__mmask8)-1, \
-                                       (int)(scale))
-
-#define _mm512_mask_i64gather_ps(v1_old, mask, index, addr, scale) \
-  (__m256)__builtin_ia32_gatherdiv16sf((__v8sf)(__m256)(v1_old),\
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), \
-                                       (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i64gather_epi32(index, addr, scale) \
-  (__m256i)__builtin_ia32_gatherdiv16si((__v8si)_mm256_undefined_si256(), \
-                                        (void const *)(addr), \
-                                        (__v8di)(__m512i)(index), \
-                                        (__mmask8)-1, (int)(scale))
-
-#define _mm512_mask_i64gather_epi32(v1_old, mask, index, addr, scale) \
-  (__m256i)__builtin_ia32_gatherdiv16si((__v8si)(__m256i)(v1_old), \
-                                        (void const *)(addr), \
-                                        (__v8di)(__m512i)(index), \
-                                        (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i64gather_pd(index, addr, scale) \
-  (__m512d)__builtin_ia32_gatherdiv8df((__v8df)_mm512_undefined_pd(), \
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), (__mmask8)-1, \
-                                       (int)(scale))
-
-#define _mm512_mask_i64gather_pd(v1_old, mask, index, addr, scale) \
-  (__m512d)__builtin_ia32_gatherdiv8df((__v8df)(__m512d)(v1_old), \
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), \
-                                       (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i64gather_epi64(index, addr, scale) \
-  (__m512i)__builtin_ia32_gatherdiv8di((__v8di)_mm512_undefined_epi32(), \
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), (__mmask8)-1, \
-                                       (int)(scale))
-
-#define _mm512_mask_i64gather_epi64(v1_old, mask, index, addr, scale) \
-  (__m512i)__builtin_ia32_gatherdiv8di((__v8di)(__m512i)(v1_old), \
-                                       (void const *)(addr), \
-                                       (__v8di)(__m512i)(index), \
-                                       (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i32gather_ps(index, addr, scale) \
-  (__m512)__builtin_ia32_gathersiv16sf((__v16sf)_mm512_undefined_ps(), \
-                                       (void const *)(addr), \
-                                       (__v16sf)(__m512)(index), \
-                                       (__mmask16)-1, (int)(scale))
-
-#define _mm512_mask_i32gather_ps(v1_old, mask, index, addr, scale) \
-  (__m512)__builtin_ia32_gathersiv16sf((__v16sf)(__m512)(v1_old), \
-                                       (void const *)(addr), \
-                                       (__v16sf)(__m512)(index), \
-                                       (__mmask16)(mask), (int)(scale))
-
-#define _mm512_i32gather_epi32(index, addr, scale) \
-  (__m512i)__builtin_ia32_gathersiv16si((__v16si)_mm512_undefined_epi32(), \
-                                        (void const *)(addr), \
-                                        (__v16si)(__m512i)(index), \
-                                        (__mmask16)-1, (int)(scale))
-
-#define _mm512_mask_i32gather_epi32(v1_old, mask, index, addr, scale) \
-  (__m512i)__builtin_ia32_gathersiv16si((__v16si)(__m512i)(v1_old), \
-                                        (void const *)(addr), \
-                                        (__v16si)(__m512i)(index), \
-                                        (__mmask16)(mask), (int)(scale))
-
-#define _mm512_i32gather_pd(index, addr, scale) \
-  (__m512d)__builtin_ia32_gathersiv8df((__v8df)_mm512_undefined_pd(), \
-                                       (void const *)(addr), \
-                                       (__v8si)(__m256i)(index), (__mmask8)-1, \
-                                       (int)(scale))
-
-#define _mm512_mask_i32gather_pd(v1_old, mask, index, addr, scale) \
-  (__m512d)__builtin_ia32_gathersiv8df((__v8df)(__m512d)(v1_old), \
-                                       (void const *)(addr), \
-                                       (__v8si)(__m256i)(index), \
-                                       (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i32gather_epi64(index, addr, scale) \
-  (__m512i)__builtin_ia32_gathersiv8di((__v8di)_mm512_undefined_epi32(), \
-                                       (void const *)(addr), \
-                                       (__v8si)(__m256i)(index), (__mmask8)-1, \
-                                       (int)(scale))
-
-#define _mm512_mask_i32gather_epi64(v1_old, mask, index, addr, scale) \
-  (__m512i)__builtin_ia32_gathersiv8di((__v8di)(__m512i)(v1_old), \
-                                       (void const *)(addr), \
-                                       (__v8si)(__m256i)(index), \
-                                       (__mmask8)(mask), (int)(scale))
-
-#define _mm512_i64scatter_ps(addr, index, v1, scale) \
-  __builtin_ia32_scatterdiv16sf((void *)(addr), (__mmask8)-1, \
-                                (__v8di)(__m512i)(index), \
-                                (__v8sf)(__m256)(v1), (int)(scale))
-
-#define _mm512_mask_i64scatter_ps(addr, mask, index, v1, scale) \
-  __builtin_ia32_scatterdiv16sf((void *)(addr), (__mmask8)(mask), \
-                                (__v8di)(__m512i)(index), \
-                                (__v8sf)(__m256)(v1), (int)(scale))
-
-#define _mm512_i64scatter_epi32(addr, index, v1, scale) \
-  __builtin_ia32_scatterdiv16si((void *)(addr), (__mmask8)-1, \
-                                (__v8di)(__m512i)(index), \
-                                (__v8si)(__m256i)(v1), (int)(scale))
-
-#define _mm512_mask_i64scatter_epi32(addr, mask, index, v1, scale) \
-  __builtin_ia32_scatterdiv16si((void *)(addr), (__mmask8)(mask), \
-                                (__v8di)(__m512i)(index), \
-                                (__v8si)(__m256i)(v1), (int)(scale))
-
-#define _mm512_i64scatter_pd(addr, index, v1, scale) \
-  __builtin_ia32_scatterdiv8df((void *)(addr), (__mmask8)-1, \
-                               (__v8di)(__m512i)(index), \
-                               (__v8df)(__m512d)(v1), (int)(scale))
-
-#define _mm512_mask_i64scatter_pd(addr, mask, index, v1, scale) \
-  __builtin_ia32_scatterdiv8df((void *)(addr), (__mmask8)(mask), \
-                               (__v8di)(__m512i)(index), \
-                               (__v8df)(__m512d)(v1), (int)(scale))
-
-#define _mm512_i64scatter_epi64(addr, index, v1, scale) \
-  __builtin_ia32_scatterdiv8di((void *)(addr), (__mmask8)-1, \
-                               (__v8di)(__m512i)(index), \
-                               (__v8di)(__m512i)(v1), (int)(scale))
-
-#define _mm512_mask_i64scatter_epi64(addr, mask, index, v1, scale) \
-  __builtin_ia32_scatterdiv8di((void *)(addr), (__mmask8)(mask), \
-                               (__v8di)(__m512i)(index), \
-                               (__v8di)(__m512i)(v1), (int)(scale))
-
-#define _mm512_i32scatter_ps(addr, index, v1, scale) \
-  __builtin_ia32_scattersiv16sf((void *)(addr), (__mmask16)-1, \
-                                (__v16si)(__m512i)(index), \
-                                (__v16sf)(__m512)(v1), (int)(scale))
-
-#define _mm512_mask_i32scatter_ps(addr, mask, index, v1, scale) \
-  __builtin_ia32_scattersiv16sf((void *)(addr), (__mmask16)(mask), \
-                                (__v16si)(__m512i)(index), \
-                                (__v16sf)(__m512)(v1), (int)(scale))
-
-#define _mm512_i32scatter_epi32(addr, index, v1, scale) \
-  __builtin_ia32_scattersiv16si((void *)(addr), (__mmask16)-1, \
-                                (__v16si)(__m512i)(index), \
-                                (__v16si)(__m512i)(v1), (int)(scale))
-
-#define _mm512_mask_i32scatter_epi32(addr, mask, index, v1, scale) \
-  __builtin_ia32_scattersiv16si((void *)(addr), (__mmask16)(mask), \
-                                (__v16si)(__m512i)(index), \
-                                (__v16si)(__m512i)(v1), (int)(scale))
-
-#define _mm512_i32scatter_pd(addr, index, v1, scale) \
-  __builtin_ia32_scattersiv8df((void *)(addr), (__mmask8)-1, \
-                               (__v8si)(__m256i)(index), \
-                               (__v8df)(__m512d)(v1), (int)(scale))
-
-#define _mm512_mask_i32scatter_pd(addr, mask, index, v1, scale) \
-  __builtin_ia32_scattersiv8df((void *)(addr), (__mmask8)(mask), \
-                               (__v8si)(__m256i)(index), \
-                               (__v8df)(__m512d)(v1), (int)(scale))
-
-#define _mm512_i32scatter_epi64(addr, index, v1, scale) \
-  __builtin_ia32_scattersiv8di((void *)(addr), (__mmask8)-1, \
-                               (__v8si)(__m256i)(index), \
-                               (__v8di)(__m512i)(v1), (int)(scale))
-
-#define _mm512_mask_i32scatter_epi64(addr, mask, index, v1, scale) \
-  __builtin_ia32_scattersiv8di((void *)(addr), (__mmask8)(mask), \
-                               (__v8si)(__m256i)(index), \
-                               (__v8di)(__m512i)(v1), (int)(scale))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_fmadd_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_vfmaddss3_mask((__v4sf)__W,
-                                       (__v4sf)__A,
-                                       (__v4sf)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fmadd_round_ss(A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(A), \
-                                        (__v4sf)(__m128)(B), \
-                                        (__v4sf)(__m128)(C), (__mmask8)-1, \
-                                        (int)(R))
-
-#define _mm_mask_fmadd_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(W), \
-                                        (__v4sf)(__m128)(A), \
-                                        (__v4sf)(__m128)(B), (__mmask8)(U), \
-                                        (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_fmadd_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C)
-{
-  return __builtin_ia32_vfmaddss3_maskz((__v4sf)__A,
-                                        (__v4sf)__B,
-                                        (__v4sf)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fmadd_round_ss(U, A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_maskz((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         (__v4sf)(__m128)(C), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask3_fmadd_ss (__m128 __W, __m128 __X, __m128 __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmaddss3_mask3((__v4sf)__W,
-                                        (__v4sf)__X,
-                                        (__v4sf)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fmadd_round_ss(W, X, Y, U, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask3((__v4sf)(__m128)(W), \
-                                         (__v4sf)(__m128)(X), \
-                                         (__v4sf)(__m128)(Y), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_fmsub_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_vfmaddss3_mask((__v4sf)__W,
-                                       (__v4sf)__A,
-                                       -(__v4sf)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fmsub_round_ss(A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(A), \
-                                        (__v4sf)(__m128)(B), \
-                                        -(__v4sf)(__m128)(C), (__mmask8)-1, \
-                                        (int)(R))
-
-#define _mm_mask_fmsub_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(W), \
-                                        (__v4sf)(__m128)(A), \
-                                        -(__v4sf)(__m128)(B), (__mmask8)(U), \
-                                        (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_fmsub_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C)
-{
-  return __builtin_ia32_vfmaddss3_maskz((__v4sf)__A,
-                                        (__v4sf)__B,
-                                        -(__v4sf)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fmsub_round_ss(U, A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_maskz((__v4sf)(__m128)(A), \
-                                         (__v4sf)(__m128)(B), \
-                                         -(__v4sf)(__m128)(C), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask3_fmsub_ss (__m128 __W, __m128 __X, __m128 __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmsubss3_mask3((__v4sf)__W,
-                                        (__v4sf)__X,
-                                        (__v4sf)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fmsub_round_ss(W, X, Y, U, R) \
-  (__m128)__builtin_ia32_vfmsubss3_mask3((__v4sf)(__m128)(W), \
-                                         (__v4sf)(__m128)(X), \
-                                         (__v4sf)(__m128)(Y), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_fnmadd_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_vfmaddss3_mask((__v4sf)__W,
-                                       -(__v4sf)__A,
-                                       (__v4sf)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fnmadd_round_ss(A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(A), \
-                                        -(__v4sf)(__m128)(B), \
-                                        (__v4sf)(__m128)(C), (__mmask8)-1, \
-                                        (int)(R))
-
-#define _mm_mask_fnmadd_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(W), \
-                                        -(__v4sf)(__m128)(A), \
-                                        (__v4sf)(__m128)(B), (__mmask8)(U), \
-                                        (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_fnmadd_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C)
-{
-  return __builtin_ia32_vfmaddss3_maskz((__v4sf)__A,
-                                        -(__v4sf)__B,
-                                        (__v4sf)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fnmadd_round_ss(U, A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_maskz((__v4sf)(__m128)(A), \
-                                         -(__v4sf)(__m128)(B), \
-                                         (__v4sf)(__m128)(C), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask3_fnmadd_ss (__m128 __W, __m128 __X, __m128 __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmaddss3_mask3((__v4sf)__W,
-                                        -(__v4sf)__X,
-                                        (__v4sf)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fnmadd_round_ss(W, X, Y, U, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask3((__v4sf)(__m128)(W), \
-                                         -(__v4sf)(__m128)(X), \
-                                         (__v4sf)(__m128)(Y), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_fnmsub_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_vfmaddss3_mask((__v4sf)__W,
-                                       -(__v4sf)__A,
-                                       -(__v4sf)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fnmsub_round_ss(A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(A), \
-                                        -(__v4sf)(__m128)(B), \
-                                        -(__v4sf)(__m128)(C), (__mmask8)-1, \
-                                        (int)(R))
-
-#define _mm_mask_fnmsub_round_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_vfmaddss3_mask((__v4sf)(__m128)(W), \
-                                        -(__v4sf)(__m128)(A), \
-                                        -(__v4sf)(__m128)(B), (__mmask8)(U), \
-                                        (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_fnmsub_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C)
-{
-  return __builtin_ia32_vfmaddss3_maskz((__v4sf)__A,
-                                        -(__v4sf)__B,
-                                        -(__v4sf)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fnmsub_round_ss(U, A, B, C, R) \
-  (__m128)__builtin_ia32_vfmaddss3_maskz((__v4sf)(__m128)(A), \
-                                         -(__v4sf)(__m128)(B), \
-                                         -(__v4sf)(__m128)(C), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask3_fnmsub_ss (__m128 __W, __m128 __X, __m128 __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmsubss3_mask3((__v4sf)__W,
-                                        -(__v4sf)__X,
-                                        (__v4sf)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fnmsub_round_ss(W, X, Y, U, R) \
-  (__m128)__builtin_ia32_vfmsubss3_mask3((__v4sf)(__m128)(W), \
-                                         -(__v4sf)(__m128)(X), \
-                                         (__v4sf)(__m128)(Y), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_fmadd_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_vfmaddsd3_mask((__v2df)__W,
-                                       (__v2df)__A,
-                                       (__v2df)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fmadd_round_sd(A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(A), \
-                                         (__v2df)(__m128d)(B), \
-                                         (__v2df)(__m128d)(C), (__mmask8)-1, \
-                                         (int)(R))
-
-#define _mm_mask_fmadd_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(W), \
-                                         (__v2df)(__m128d)(A), \
-                                         (__v2df)(__m128d)(B), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_fmadd_sd (__mmask8 __U, __m128d __A, __m128d __B, __m128d __C)
-{
-  return __builtin_ia32_vfmaddsd3_maskz((__v2df)__A,
-                                        (__v2df)__B,
-                                        (__v2df)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fmadd_round_sd(U, A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_maskz((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          (__v2df)(__m128d)(C), (__mmask8)(U), \
-                                          (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask3_fmadd_sd (__m128d __W, __m128d __X, __m128d __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmaddsd3_mask3((__v2df)__W,
-                                        (__v2df)__X,
-                                        (__v2df)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fmadd_round_sd(W, X, Y, U, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask3((__v2df)(__m128d)(W), \
-                                          (__v2df)(__m128d)(X), \
-                                          (__v2df)(__m128d)(Y), (__mmask8)(U), \
-                                          (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_fmsub_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_vfmaddsd3_mask((__v2df)__W,
-                                       (__v2df)__A,
-                                       -(__v2df)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fmsub_round_sd(A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(A), \
-                                         (__v2df)(__m128d)(B), \
-                                         -(__v2df)(__m128d)(C), (__mmask8)-1, \
-                                         (int)(R))
-
-#define _mm_mask_fmsub_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(W), \
-                                         (__v2df)(__m128d)(A), \
-                                         -(__v2df)(__m128d)(B), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_fmsub_sd (__mmask8 __U, __m128d __A, __m128d __B, __m128d __C)
-{
-  return __builtin_ia32_vfmaddsd3_maskz((__v2df)__A,
-                                        (__v2df)__B,
-                                        -(__v2df)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fmsub_round_sd(U, A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_maskz((__v2df)(__m128d)(A), \
-                                          (__v2df)(__m128d)(B), \
-                                          -(__v2df)(__m128d)(C), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask3_fmsub_sd (__m128d __W, __m128d __X, __m128d __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmsubsd3_mask3((__v2df)__W,
-                                        (__v2df)__X,
-                                        (__v2df)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fmsub_round_sd(W, X, Y, U, R) \
-  (__m128d)__builtin_ia32_vfmsubsd3_mask3((__v2df)(__m128d)(W), \
-                                          (__v2df)(__m128d)(X), \
-                                          (__v2df)(__m128d)(Y), \
-                                          (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_fnmadd_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_vfmaddsd3_mask((__v2df)__W,
-                                       -(__v2df)__A,
-                                       (__v2df)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fnmadd_round_sd(A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(A), \
-                                         -(__v2df)(__m128d)(B), \
-                                         (__v2df)(__m128d)(C), (__mmask8)-1, \
-                                         (int)(R))
-
-#define _mm_mask_fnmadd_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(W), \
-                                         -(__v2df)(__m128d)(A), \
-                                         (__v2df)(__m128d)(B), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_fnmadd_sd (__mmask8 __U, __m128d __A, __m128d __B, __m128d __C)
-{
-  return __builtin_ia32_vfmaddsd3_maskz((__v2df)__A,
-                                        -(__v2df)__B,
-                                        (__v2df)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fnmadd_round_sd(U, A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_maskz((__v2df)(__m128d)(A), \
-                                          -(__v2df)(__m128d)(B), \
-                                          (__v2df)(__m128d)(C), (__mmask8)(U), \
-                                          (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask3_fnmadd_sd (__m128d __W, __m128d __X, __m128d __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmaddsd3_mask3((__v2df)__W,
-                                        -(__v2df)__X,
-                                        (__v2df)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fnmadd_round_sd(W, X, Y, U, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask3((__v2df)(__m128d)(W), \
-                                          -(__v2df)(__m128d)(X), \
-                                          (__v2df)(__m128d)(Y), (__mmask8)(U), \
-                                          (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_fnmsub_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_vfmaddsd3_mask((__v2df)__W,
-                                       -(__v2df)__A,
-                                       -(__v2df)__B,
-                                       (__mmask8)__U,
-                                       _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_fnmsub_round_sd(A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(A), \
-                                         -(__v2df)(__m128d)(B), \
-                                         -(__v2df)(__m128d)(C), (__mmask8)-1, \
-                                         (int)(R))
-
-#define _mm_mask_fnmsub_round_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_mask((__v2df)(__m128d)(W), \
-                                         -(__v2df)(__m128d)(A), \
-                                         -(__v2df)(__m128d)(B), (__mmask8)(U), \
-                                         (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_fnmsub_sd (__mmask8 __U, __m128d __A, __m128d __B, __m128d __C)
-{
-  return __builtin_ia32_vfmaddsd3_maskz((__v2df)__A,
-                                        -(__v2df)__B,
-                                        -(__v2df)__C,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_maskz_fnmsub_round_sd(U, A, B, C, R) \
-  (__m128d)__builtin_ia32_vfmaddsd3_maskz((__v2df)(__m128d)(A), \
-                                          -(__v2df)(__m128d)(B), \
-                                          -(__v2df)(__m128d)(C), \
-                                          (__mmask8)(U), \
-                                          (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask3_fnmsub_sd (__m128d __W, __m128d __X, __m128d __Y, __mmask8 __U)
-{
-  return __builtin_ia32_vfmsubsd3_mask3((__v2df)__W,
-                                        -(__v2df)__X,
-                                        (__v2df)__Y,
-                                        (__mmask8)__U,
-                                        _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_mask3_fnmsub_round_sd(W, X, Y, U, R) \
-  (__m128d)__builtin_ia32_vfmsubsd3_mask3((__v2df)(__m128d)(W), \
-                                          -(__v2df)(__m128d)(X), \
-                                          (__v2df)(__m128d)(Y), \
-                                          (__mmask8)(U), (int)(R))
-
-#define _mm512_permutex_pd(X, C) \
-  (__m512d)__builtin_ia32_permdf512((__v8df)(__m512d)(X), (int)(C))
-
-#define _mm512_mask_permutex_pd(W, U, X, C) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_permutex_pd((X), (C)), \
-                                       (__v8df)(__m512d)(W))
-
-#define _mm512_maskz_permutex_pd(U, X, C) \
-  (__m512d)__builtin_ia32_selectpd_512((__mmask8)(U), \
-                                       (__v8df)_mm512_permutex_pd((X), (C)), \
-                                       (__v8df)_mm512_setzero_pd())
-
-#define _mm512_permutex_epi64(X, C) \
-  (__m512i)__builtin_ia32_permdi512((__v8di)(__m512i)(X), (int)(C))
-
-#define _mm512_mask_permutex_epi64(W, U, X, C) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_permutex_epi64((X), (C)), \
-                                      (__v8di)(__m512i)(W))
-
-#define _mm512_maskz_permutex_epi64(U, X, C) \
-  (__m512i)__builtin_ia32_selectq_512((__mmask8)(U), \
-                                      (__v8di)_mm512_permutex_epi64((X), (C)), \
-                                      (__v8di)_mm512_setzero_si512())
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_permutexvar_pd (__m512i __X, __m512d __Y)
-{
-  return (__m512d)__builtin_ia32_permvardf512((__v8df) __Y, (__v8di) __X);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_permutexvar_pd (__m512d __W, __mmask8 __U, __m512i __X, __m512d __Y)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                        (__v8df)_mm512_permutexvar_pd(__X, __Y),
-                                        (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutexvar_pd (__mmask8 __U, __m512i __X, __m512d __Y)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                        (__v8df)_mm512_permutexvar_pd(__X, __Y),
-                                        (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_permutexvar_epi64 (__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_permvardi512((__v8di)__Y, (__v8di)__X);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutexvar_epi64 (__mmask8 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                     (__v8di)_mm512_permutexvar_epi64(__X, __Y),
-                                     (__v8di)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_permutexvar_epi64 (__m512i __W, __mmask8 __M, __m512i __X,
-             __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M,
-                                     (__v8di)_mm512_permutexvar_epi64(__X, __Y),
-                                     (__v8di)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_permutexvar_ps (__m512i __X, __m512 __Y)
-{
-  return (__m512)__builtin_ia32_permvarsf512((__v16sf)__Y, (__v16si)__X);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_permutexvar_ps (__m512 __W, __mmask16 __U, __m512i __X, __m512 __Y)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                       (__v16sf)_mm512_permutexvar_ps(__X, __Y),
-                                       (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutexvar_ps (__mmask16 __U, __m512i __X, __m512 __Y)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                       (__v16sf)_mm512_permutexvar_ps(__X, __Y),
-                                       (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_permutexvar_epi32 (__m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_permvarsi512((__v16si)__Y, (__v16si)__X);
-}
-
-#define _mm512_permutevar_epi32 _mm512_permutexvar_epi32
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_permutexvar_epi32 (__mmask16 __M, __m512i __X, __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                    (__v16si)_mm512_permutexvar_epi32(__X, __Y),
-                                    (__v16si)_mm512_setzero_si512());
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_permutexvar_epi32 (__m512i __W, __mmask16 __M, __m512i __X,
-             __m512i __Y)
-{
-  return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M,
-                                    (__v16si)_mm512_permutexvar_epi32(__X, __Y),
-                                    (__v16si)__W);
-}
-
-#define _mm512_mask_permutevar_epi32 _mm512_mask_permutexvar_epi32
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kand (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_kandhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kandn (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_kandnhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kor (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_korhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm512_kortestc (__mmask16 __A, __mmask16 __B)
-{
-  return __builtin_ia32_kortestchi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm512_kortestz (__mmask16 __A, __mmask16 __B)
-{
-  return __builtin_ia32_kortestzhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ unsigned char __DEFAULT_FN_ATTRS
-_kortestc_mask16_u8(__mmask16 __A, __mmask16 __B)
-{
-  return (unsigned char)__builtin_ia32_kortestchi(__A, __B);
-}
-
-static __inline__ unsigned char __DEFAULT_FN_ATTRS
-_kortestz_mask16_u8(__mmask16 __A, __mmask16 __B)
-{
-  return (unsigned char)__builtin_ia32_kortestzhi(__A, __B);
-}
-
-static __inline__ unsigned char __DEFAULT_FN_ATTRS
-_kortest_mask16_u8(__mmask16 __A, __mmask16 __B, unsigned char *__C) {
-  *__C = (unsigned char)__builtin_ia32_kortestchi(__A, __B);
-  return (unsigned char)__builtin_ia32_kortestzhi(__A, __B);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kunpackb (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_kunpckhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kxnor (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_kxnorhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_mm512_kxor (__mmask16 __A, __mmask16 __B)
-{
-  return (__mmask16) __builtin_ia32_kxorhi ((__mmask16) __A, (__mmask16) __B);
-}
-
-#define _kand_mask16 _mm512_kand
-#define _kandn_mask16 _mm512_kandn
-#define _knot_mask16 _mm512_knot
-#define _kor_mask16 _mm512_kor
-#define _kxnor_mask16 _mm512_kxnor
-#define _kxor_mask16 _mm512_kxor
-
-#define _kshiftli_mask16(A, I) \
-  (__mmask16)__builtin_ia32_kshiftlihi((__mmask16)(A), (unsigned int)(I))
-
-#define _kshiftri_mask16(A, I) \
-  (__mmask16)__builtin_ia32_kshiftrihi((__mmask16)(A), (unsigned int)(I))
-
-static __inline__ unsigned int __DEFAULT_FN_ATTRS
-_cvtmask16_u32(__mmask16 __A) {
-  return (unsigned int)__builtin_ia32_kmovw((__mmask16)__A);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_cvtu32_mask16(unsigned int __A) {
-  return (__mmask16)__builtin_ia32_kmovw((__mmask16)__A);
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS
-_load_mask16(__mmask16 *__A) {
-  return (__mmask16)__builtin_ia32_kmovw(*(__mmask16 *)__A);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS
-_store_mask16(__mmask16 *__A, __mmask16 __B) {
-  *(__mmask16 *)__A = __builtin_ia32_kmovw((__mmask16)__B);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_si512 (__m512i * __P, __m512i __A)
-{
-  typedef __v8di __v8di_aligned __attribute__((aligned(64)));
-  __builtin_nontemporal_store((__v8di_aligned)__A, (__v8di_aligned*)__P);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_stream_load_si512 (void const *__P)
-{
-  typedef __v8di __v8di_aligned __attribute__((aligned(64)));
-  return (__m512i) __builtin_nontemporal_load((const __v8di_aligned *)__P);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_pd (double *__P, __m512d __A)
-{
-  typedef __v8df __v8df_aligned __attribute__((aligned(64)));
-  __builtin_nontemporal_store((__v8df_aligned)__A, (__v8df_aligned*)__P);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_stream_ps (float *__P, __m512 __A)
-{
-  typedef __v16sf __v16sf_aligned __attribute__((aligned(64)));
-  __builtin_nontemporal_store((__v16sf_aligned)__A, (__v16sf_aligned*)__P);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_compress_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_compressdf512_mask ((__v8df) __A,
-                  (__v8df) __W,
-                  (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_compress_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_compressdf512_mask ((__v8df) __A,
-                  (__v8df)
-                  _mm512_setzero_pd (),
-                  (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_compress_epi64 (__m512i __W, __mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_compressdi512_mask ((__v8di) __A,
-                  (__v8di) __W,
-                  (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_compress_epi64 (__mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_compressdi512_mask ((__v8di) __A,
-                  (__v8di)
-                  _mm512_setzero_si512 (),
-                  (__mmask8) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_compress_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_compresssf512_mask ((__v16sf) __A,
-                 (__v16sf) __W,
-                 (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_compress_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_compresssf512_mask ((__v16sf) __A,
-                 (__v16sf)
-                 _mm512_setzero_ps (),
-                 (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_compress_epi32 (__m512i __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_compresssi512_mask ((__v16si) __A,
-                  (__v16si) __W,
-                  (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_compress_epi32 (__mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_compresssi512_mask ((__v16si) __A,
-                  (__v16si)
-                  _mm512_setzero_si512 (),
-                  (__mmask16) __U);
-}
-
-#define _mm_cmp_round_ss_mask(X, Y, P, R) \
-  (__mmask8)__builtin_ia32_cmpss_mask((__v4sf)(__m128)(X), \
-                                      (__v4sf)(__m128)(Y), (int)(P), \
-                                      (__mmask8)-1, (int)(R))
-
-#define _mm_mask_cmp_round_ss_mask(M, X, Y, P, R) \
-  (__mmask8)__builtin_ia32_cmpss_mask((__v4sf)(__m128)(X), \
-                                      (__v4sf)(__m128)(Y), (int)(P), \
-                                      (__mmask8)(M), (int)(R))
-
-#define _mm_cmp_ss_mask(X, Y, P) \
-  (__mmask8)__builtin_ia32_cmpss_mask((__v4sf)(__m128)(X), \
-                                      (__v4sf)(__m128)(Y), (int)(P), \
-                                      (__mmask8)-1, \
-                                      _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_cmp_ss_mask(M, X, Y, P) \
-  (__mmask8)__builtin_ia32_cmpss_mask((__v4sf)(__m128)(X), \
-                                      (__v4sf)(__m128)(Y), (int)(P), \
-                                      (__mmask8)(M), \
-                                      _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_cmp_round_sd_mask(X, Y, P, R) \
-  (__mmask8)__builtin_ia32_cmpsd_mask((__v2df)(__m128d)(X), \
-                                      (__v2df)(__m128d)(Y), (int)(P), \
-                                      (__mmask8)-1, (int)(R))
-
-#define _mm_mask_cmp_round_sd_mask(M, X, Y, P, R) \
-  (__mmask8)__builtin_ia32_cmpsd_mask((__v2df)(__m128d)(X), \
-                                      (__v2df)(__m128d)(Y), (int)(P), \
-                                      (__mmask8)(M), (int)(R))
-
-#define _mm_cmp_sd_mask(X, Y, P) \
-  (__mmask8)__builtin_ia32_cmpsd_mask((__v2df)(__m128d)(X), \
-                                      (__v2df)(__m128d)(Y), (int)(P), \
-                                      (__mmask8)-1, \
-                                      _MM_FROUND_CUR_DIRECTION)
-
-#define _mm_mask_cmp_sd_mask(M, X, Y, P) \
-  (__mmask8)__builtin_ia32_cmpsd_mask((__v2df)(__m128d)(X), \
-                                      (__v2df)(__m128d)(Y), (int)(P), \
-                                      (__mmask8)(M), \
-                                      _MM_FROUND_CUR_DIRECTION)
-
-/* Bit Test */
-
-static __inline __mmask16 __DEFAULT_FN_ATTRS512
-_mm512_test_epi32_mask (__m512i __A, __m512i __B)
-{
-  return _mm512_cmpneq_epi32_mask (_mm512_and_epi32(__A, __B),
-                                   _mm512_setzero_si512());
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS512
-_mm512_mask_test_epi32_mask (__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return _mm512_mask_cmpneq_epi32_mask (__U, _mm512_and_epi32 (__A, __B),
-                                        _mm512_setzero_si512());
-}
-
-static __inline __mmask8 __DEFAULT_FN_ATTRS512
-_mm512_test_epi64_mask (__m512i __A, __m512i __B)
-{
-  return _mm512_cmpneq_epi64_mask (_mm512_and_epi32 (__A, __B),
-                                   _mm512_setzero_si512());
-}
-
-static __inline__ __mmask8 __DEFAULT_FN_ATTRS512
-_mm512_mask_test_epi64_mask (__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return _mm512_mask_cmpneq_epi64_mask (__U, _mm512_and_epi32 (__A, __B),
-                                        _mm512_setzero_si512());
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS512
-_mm512_testn_epi32_mask (__m512i __A, __m512i __B)
-{
-  return _mm512_cmpeq_epi32_mask (_mm512_and_epi32 (__A, __B),
-                                  _mm512_setzero_si512());
-}
-
-static __inline__ __mmask16 __DEFAULT_FN_ATTRS512
-_mm512_mask_testn_epi32_mask (__mmask16 __U, __m512i __A, __m512i __B)
-{
-  return _mm512_mask_cmpeq_epi32_mask (__U, _mm512_and_epi32 (__A, __B),
-                                       _mm512_setzero_si512());
-}
-
-static __inline__ __mmask8 __DEFAULT_FN_ATTRS512
-_mm512_testn_epi64_mask (__m512i __A, __m512i __B)
-{
-  return _mm512_cmpeq_epi64_mask (_mm512_and_epi32 (__A, __B),
-                                  _mm512_setzero_si512());
-}
-
-static __inline__ __mmask8 __DEFAULT_FN_ATTRS512
-_mm512_mask_testn_epi64_mask (__mmask8 __U, __m512i __A, __m512i __B)
-{
-  return _mm512_mask_cmpeq_epi64_mask (__U, _mm512_and_epi32 (__A, __B),
-                                       _mm512_setzero_si512());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_movehdup_ps (__m512 __A)
-{
-  return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
-                         1, 1, 3, 3, 5, 5, 7, 7, 9, 9, 11, 11, 13, 13, 15, 15);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_movehdup_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_movehdup_ps(__A),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_movehdup_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_movehdup_ps(__A),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_moveldup_ps (__m512 __A)
-{
-  return (__m512)__builtin_shufflevector((__v16sf)__A, (__v16sf)__A,
-                         0, 0, 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12, 14, 14);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_moveldup_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_moveldup_ps(__A),
-                                             (__v16sf)__W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_moveldup_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512)__builtin_ia32_selectps_512((__mmask16)__U,
-                                             (__v16sf)_mm512_moveldup_ps(__A),
-                                             (__v16sf)_mm512_setzero_ps());
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_move_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_selectss_128(__U, _mm_move_ss(__A, __B), __W);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_move_ss (__mmask8 __U, __m128 __A, __m128 __B)
-{
-  return __builtin_ia32_selectss_128(__U, _mm_move_ss(__A, __B),
-                                     _mm_setzero_ps());
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_move_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_selectsd_128(__U, _mm_move_sd(__A, __B), __W);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_move_sd (__mmask8 __U, __m128d __A, __m128d __B)
-{
-  return __builtin_ia32_selectsd_128(__U, _mm_move_sd(__A, __B),
-                                     _mm_setzero_pd());
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS128
-_mm_mask_store_ss (float * __W, __mmask8 __U, __m128 __A)
-{
-  __builtin_ia32_storess128_mask ((__v4sf *)__W, __A, __U & 1);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS128
-_mm_mask_store_sd (double * __W, __mmask8 __U, __m128d __A)
-{
-  __builtin_ia32_storesd128_mask ((__v2df *)__W, __A, __U & 1);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_load_ss (__m128 __W, __mmask8 __U, const float* __A)
-{
-  __m128 src = (__v4sf) __builtin_shufflevector((__v4sf) __W,
-                                                (__v4sf)_mm_setzero_ps(),
-                                                0, 4, 4, 4);
-
-  return (__m128) __builtin_ia32_loadss128_mask ((__v4sf *) __A, src, __U & 1);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_load_ss (__mmask8 __U, const float* __A)
-{
-  return (__m128)__builtin_ia32_loadss128_mask ((__v4sf *) __A,
-                                                (__v4sf) _mm_setzero_ps(),
-                                                __U & 1);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_load_sd (__m128d __W, __mmask8 __U, const double* __A)
-{
-  __m128d src = (__v2df) __builtin_shufflevector((__v2df) __W,
-                                                 (__v2df)_mm_setzero_pd(),
-                                                 0, 2);
-
-  return (__m128d) __builtin_ia32_loadsd128_mask ((__v2df *) __A, src, __U & 1);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_load_sd (__mmask8 __U, const double* __A)
-{
-  return (__m128d) __builtin_ia32_loadsd128_mask ((__v2df *) __A,
-                                                  (__v2df) _mm_setzero_pd(),
-                                                  __U & 1);
-}
-
-#define _mm512_shuffle_epi32(A, I) \
-  (__m512i)__builtin_ia32_pshufd512((__v16si)(__m512i)(A), (int)(I))
-
-#define _mm512_mask_shuffle_epi32(W, U, A, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_shuffle_epi32((A), (I)), \
-                                      (__v16si)(__m512i)(W))
-
-#define _mm512_maskz_shuffle_epi32(U, A, I) \
-  (__m512i)__builtin_ia32_selectd_512((__mmask16)(U), \
-                                      (__v16si)_mm512_shuffle_epi32((A), (I)), \
-                                      (__v16si)_mm512_setzero_si512())
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_expand_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_expanddf512_mask ((__v8df) __A,
-                (__v8df) __W,
-                (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_expand_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_expanddf512_mask ((__v8df) __A,
-                (__v8df) _mm512_setzero_pd (),
-                (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_expand_epi64 (__m512i __W, __mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_expanddi512_mask ((__v8di) __A,
-                (__v8di) __W,
-                (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_expand_epi64 ( __mmask8 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_expanddi512_mask ((__v8di) __A,
-                (__v8di) _mm512_setzero_si512 (),
-                (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_expandloadu_pd(__m512d __W, __mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_expandloaddf512_mask ((const __v8df *)__P,
-              (__v8df) __W,
-              (__mmask8) __U);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_expandloadu_pd(__mmask8 __U, void const *__P)
-{
-  return (__m512d) __builtin_ia32_expandloaddf512_mask ((const __v8df *)__P,
-              (__v8df) _mm512_setzero_pd(),
-              (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_expandloadu_epi64(__m512i __W, __mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_expandloaddi512_mask ((const __v8di *)__P,
-              (__v8di) __W,
-              (__mmask8) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_expandloadu_epi64(__mmask8 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_expandloaddi512_mask ((const __v8di *)__P,
-              (__v8di) _mm512_setzero_si512(),
-              (__mmask8) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_expandloadu_ps(__m512 __W, __mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_expandloadsf512_mask ((const __v16sf *)__P,
-                   (__v16sf) __W,
-                   (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_expandloadu_ps(__mmask16 __U, void const *__P)
-{
-  return (__m512) __builtin_ia32_expandloadsf512_mask ((const __v16sf *)__P,
-                   (__v16sf) _mm512_setzero_ps(),
-                   (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_expandloadu_epi32(__m512i __W, __mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_expandloadsi512_mask ((const __v16si *)__P,
-              (__v16si) __W,
-              (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_expandloadu_epi32(__mmask16 __U, void const *__P)
-{
-  return (__m512i) __builtin_ia32_expandloadsi512_mask ((const __v16si *)__P,
-              (__v16si) _mm512_setzero_si512(),
-              (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_expand_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_expandsf512_mask ((__v16sf) __A,
-               (__v16sf) __W,
-               (__mmask16) __U);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_expand_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_expandsf512_mask ((__v16sf) __A,
-               (__v16sf) _mm512_setzero_ps(),
-               (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_expand_epi32 (__m512i __W, __mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_expandsi512_mask ((__v16si) __A,
-                (__v16si) __W,
-                (__mmask16) __U);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_maskz_expand_epi32 (__mmask16 __U, __m512i __A)
-{
-  return (__m512i) __builtin_ia32_expandsi512_mask ((__v16si) __A,
-                (__v16si) _mm512_setzero_si512(),
-                (__mmask16) __U);
-}
-
-#define _mm512_cvt_roundps_pd(A, R) \
-  (__m512d)__builtin_ia32_cvtps2pd512_mask((__v8sf)(__m256)(A), \
-                                           (__v8df)_mm512_undefined_pd(), \
-                                           (__mmask8)-1, (int)(R))
-
-#define _mm512_mask_cvt_roundps_pd(W, U, A, R) \
-  (__m512d)__builtin_ia32_cvtps2pd512_mask((__v8sf)(__m256)(A), \
-                                           (__v8df)(__m512d)(W), \
-                                           (__mmask8)(U), (int)(R))
-
-#define _mm512_maskz_cvt_roundps_pd(U, A, R) \
-  (__m512d)__builtin_ia32_cvtps2pd512_mask((__v8sf)(__m256)(A), \
-                                           (__v8df)_mm512_setzero_pd(), \
-                                           (__mmask8)(U), (int)(R))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtps_pd (__m256 __A)
-{
-  return (__m512d) __builtin_convertvector((__v8sf)__A, __v8df);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtps_pd (__m512d __W, __mmask8 __U, __m256 __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_cvtps_pd(__A),
-                                              (__v8df)__W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_cvtps_pd (__mmask8 __U, __m256 __A)
-{
-  return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__U,
-                                              (__v8df)_mm512_cvtps_pd(__A),
-                                              (__v8df)_mm512_setzero_pd());
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_cvtpslo_pd (__m512 __A)
-{
-  return (__m512d) _mm512_cvtps_pd(_mm512_castps512_ps256(__A));
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_cvtpslo_pd (__m512d __W, __mmask8 __U, __m512 __A)
-{
-  return (__m512d) _mm512_mask_cvtps_pd(__W, __U, _mm512_castps512_ps256(__A));
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_mov_pd (__m512d __W, __mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_selectpd_512 ((__mmask8) __U,
-              (__v8df) __A,
-              (__v8df) __W);
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_maskz_mov_pd (__mmask8 __U, __m512d __A)
-{
-  return (__m512d) __builtin_ia32_selectpd_512 ((__mmask8) __U,
-              (__v8df) __A,
-              (__v8df) _mm512_setzero_pd ());
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_mov_ps (__m512 __W, __mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_selectps_512 ((__mmask16) __U,
-             (__v16sf) __A,
-             (__v16sf) __W);
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_maskz_mov_ps (__mmask16 __U, __m512 __A)
-{
-  return (__m512) __builtin_ia32_selectps_512 ((__mmask16) __U,
-             (__v16sf) __A,
-             (__v16sf) _mm512_setzero_ps ());
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_compressstoreu_pd (void *__P, __mmask8 __U, __m512d __A)
-{
-  __builtin_ia32_compressstoredf512_mask ((__v8df *) __P, (__v8df) __A,
-            (__mmask8) __U);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_compressstoreu_epi64 (void *__P, __mmask8 __U, __m512i __A)
-{
-  __builtin_ia32_compressstoredi512_mask ((__v8di *) __P, (__v8di) __A,
-            (__mmask8) __U);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_compressstoreu_ps (void *__P, __mmask16 __U, __m512 __A)
-{
-  __builtin_ia32_compressstoresf512_mask ((__v16sf *) __P, (__v16sf) __A,
-            (__mmask16) __U);
-}
-
-static __inline__ void __DEFAULT_FN_ATTRS512
-_mm512_mask_compressstoreu_epi32 (void *__P, __mmask16 __U, __m512i __A)
-{
-  __builtin_ia32_compressstoresi512_mask ((__v16si *) __P, (__v16si) __A,
-            (__mmask16) __U);
-}
-
-#define _mm_cvt_roundsd_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtsd2ss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v2df)(__m128d)(B), \
-                                             (__v4sf)_mm_undefined_ps(), \
-                                             (__mmask8)-1, (int)(R))
-
-#define _mm_mask_cvt_roundsd_ss(W, U, A, B, R) \
-  (__m128)__builtin_ia32_cvtsd2ss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v2df)(__m128d)(B), \
-                                             (__v4sf)(__m128)(W), \
-                                             (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_cvt_roundsd_ss(U, A, B, R) \
-  (__m128)__builtin_ia32_cvtsd2ss_round_mask((__v4sf)(__m128)(A), \
-                                             (__v2df)(__m128d)(B), \
-                                             (__v4sf)_mm_setzero_ps(), \
-                                             (__mmask8)(U), (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_mask_cvtsd_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128d __B)
-{
-  return __builtin_ia32_cvtsd2ss_round_mask ((__v4sf)__A,
-                                             (__v2df)__B,
-                                             (__v4sf)__W,
-                                             (__mmask8)__U, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_maskz_cvtsd_ss (__mmask8 __U, __m128 __A, __m128d __B)
-{
-  return __builtin_ia32_cvtsd2ss_round_mask ((__v4sf)__A,
-                                             (__v2df)__B,
-                                             (__v4sf)_mm_setzero_ps(),
-                                             (__mmask8)__U, _MM_FROUND_CUR_DIRECTION);
-}
-
-#define _mm_cvtss_i32 _mm_cvtss_si32
-#define _mm_cvtsd_i32 _mm_cvtsd_si32
-#define _mm_cvti32_sd _mm_cvtsi32_sd
-#define _mm_cvti32_ss _mm_cvtsi32_ss
-#ifdef __x86_64__
-#define _mm_cvtss_i64 _mm_cvtss_si64
-#define _mm_cvtsd_i64 _mm_cvtsd_si64
-#define _mm_cvti64_sd _mm_cvtsi64_sd
-#define _mm_cvti64_ss _mm_cvtsi64_ss
-#endif
-
-#ifdef __x86_64__
-#define _mm_cvt_roundi64_sd(A, B, R) \
-  (__m128d)__builtin_ia32_cvtsi2sd64((__v2df)(__m128d)(A), (long long)(B), \
-                                     (int)(R))
-
-#define _mm_cvt_roundsi64_sd(A, B, R) \
-  (__m128d)__builtin_ia32_cvtsi2sd64((__v2df)(__m128d)(A), (long long)(B), \
-                                     (int)(R))
-#endif
-
-#define _mm_cvt_roundsi32_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtsi2ss32((__v4sf)(__m128)(A), (int)(B), (int)(R))
-
-#define _mm_cvt_roundi32_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtsi2ss32((__v4sf)(__m128)(A), (int)(B), (int)(R))
-
-#ifdef __x86_64__
-#define _mm_cvt_roundsi64_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtsi2ss64((__v4sf)(__m128)(A), (long long)(B), \
-                                    (int)(R))
-
-#define _mm_cvt_roundi64_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtsi2ss64((__v4sf)(__m128)(A), (long long)(B), \
-                                    (int)(R))
-#endif
-
-#define _mm_cvt_roundss_sd(A, B, R) \
-  (__m128d)__builtin_ia32_cvtss2sd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (__v2df)_mm_undefined_pd(), \
-                                              (__mmask8)-1, (int)(R))
-
-#define _mm_mask_cvt_roundss_sd(W, U, A, B, R) \
-  (__m128d)__builtin_ia32_cvtss2sd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (__v2df)(__m128d)(W), \
-                                              (__mmask8)(U), (int)(R))
-
-#define _mm_maskz_cvt_roundss_sd(U, A, B, R) \
-  (__m128d)__builtin_ia32_cvtss2sd_round_mask((__v2df)(__m128d)(A), \
-                                              (__v4sf)(__m128)(B), \
-                                              (__v2df)_mm_setzero_pd(), \
-                                              (__mmask8)(U), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_mask_cvtss_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128 __B)
-{
-  return __builtin_ia32_cvtss2sd_round_mask((__v2df)__A,
-                                            (__v4sf)__B,
-                                            (__v2df)__W,
-                                            (__mmask8)__U, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_maskz_cvtss_sd (__mmask8 __U, __m128d __A, __m128 __B)
-{
-  return __builtin_ia32_cvtss2sd_round_mask((__v2df)__A,
-                                            (__v4sf)__B,
-                                            (__v2df)_mm_setzero_pd(),
-                                            (__mmask8)__U, _MM_FROUND_CUR_DIRECTION);
-}
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_cvtu32_sd (__m128d __A, unsigned __B)
-{
-  __A[0] = __B;
-  return __A;
-}
-
-#ifdef __x86_64__
-#define _mm_cvt_roundu64_sd(A, B, R) \
-  (__m128d)__builtin_ia32_cvtusi2sd64((__v2df)(__m128d)(A), \
-                                      (unsigned long long)(B), (int)(R))
-
-static __inline__ __m128d __DEFAULT_FN_ATTRS128
-_mm_cvtu64_sd (__m128d __A, unsigned long long __B)
-{
-  __A[0] = __B;
-  return __A;
-}
-#endif
-
-#define _mm_cvt_roundu32_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtusi2ss32((__v4sf)(__m128)(A), (unsigned int)(B), \
-                                     (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_cvtu32_ss (__m128 __A, unsigned __B)
-{
-  __A[0] = __B;
-  return __A;
-}
-
-#ifdef __x86_64__
-#define _mm_cvt_roundu64_ss(A, B, R) \
-  (__m128)__builtin_ia32_cvtusi2ss64((__v4sf)(__m128)(A), \
-                                     (unsigned long long)(B), (int)(R))
-
-static __inline__ __m128 __DEFAULT_FN_ATTRS128
-_mm_cvtu64_ss (__m128 __A, unsigned long long __B)
-{
-  __A[0] = __B;
-  return __A;
-}
-#endif
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_set1_epi32 (__m512i __O, __mmask16 __M, int __A)
-{
-  return (__m512i) __builtin_ia32_selectd_512(__M,
-                                              (__v16si) _mm512_set1_epi32(__A),
-                                              (__v16si) __O);
-}
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_mask_set1_epi64 (__m512i __O, __mmask8 __M, long long __A)
-{
-  return (__m512i) __builtin_ia32_selectq_512(__M,
-                                              (__v8di) _mm512_set1_epi64(__A),
-                                              (__v8di) __O);
-}
-
-static  __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set_epi8 (char __e63, char __e62, char __e61, char __e60, char __e59,
-    char __e58, char __e57, char __e56, char __e55, char __e54, char __e53,
-    char __e52, char __e51, char __e50, char __e49, char __e48, char __e47,
-    char __e46, char __e45, char __e44, char __e43, char __e42, char __e41,
-    char __e40, char __e39, char __e38, char __e37, char __e36, char __e35,
-    char __e34, char __e33, char __e32, char __e31, char __e30, char __e29,
-    char __e28, char __e27, char __e26, char __e25, char __e24, char __e23,
-    char __e22, char __e21, char __e20, char __e19, char __e18, char __e17,
-    char __e16, char __e15, char __e14, char __e13, char __e12, char __e11,
-    char __e10, char __e9, char __e8, char __e7, char __e6, char __e5,
-    char __e4, char __e3, char __e2, char __e1, char __e0) {
-
-  return __extension__ (__m512i)(__v64qi)
-    {__e0, __e1, __e2, __e3, __e4, __e5, __e6, __e7,
-     __e8, __e9, __e10, __e11, __e12, __e13, __e14, __e15,
-     __e16, __e17, __e18, __e19, __e20, __e21, __e22, __e23,
-     __e24, __e25, __e26, __e27, __e28, __e29, __e30, __e31,
-     __e32, __e33, __e34, __e35, __e36, __e37, __e38, __e39,
-     __e40, __e41, __e42, __e43, __e44, __e45, __e46, __e47,
-     __e48, __e49, __e50, __e51, __e52, __e53, __e54, __e55,
-     __e56, __e57, __e58, __e59, __e60, __e61, __e62, __e63};
-}
-
-static  __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set_epi16(short __e31, short __e30, short __e29, short __e28,
-    short __e27, short __e26, short __e25, short __e24, short __e23,
-    short __e22, short __e21, short __e20, short __e19, short __e18,
-    short __e17, short __e16, short __e15, short __e14, short __e13,
-    short __e12, short __e11, short __e10, short __e9, short __e8,
-    short __e7, short __e6, short __e5, short __e4, short __e3,
-    short __e2, short __e1, short __e0) {
-  return __extension__ (__m512i)(__v32hi)
-    {__e0, __e1, __e2, __e3, __e4, __e5, __e6, __e7,
-     __e8, __e9, __e10, __e11, __e12, __e13, __e14, __e15,
-     __e16, __e17, __e18, __e19, __e20, __e21, __e22, __e23,
-     __e24, __e25, __e26, __e27, __e28, __e29, __e30, __e31 };
-}
-
-static __inline __m512i __DEFAULT_FN_ATTRS512
-_mm512_set_epi32 (int __A, int __B, int __C, int __D,
-     int __E, int __F, int __G, int __H,
-     int __I, int __J, int __K, int __L,
-     int __M, int __N, int __O, int __P)
-{
-  return __extension__ (__m512i)(__v16si)
-  { __P, __O, __N, __M, __L, __K, __J, __I,
-    __H, __G, __F, __E, __D, __C, __B, __A };
-}
-
-#define _mm512_setr_epi32(e0,e1,e2,e3,e4,e5,e6,e7,           \
-       e8,e9,e10,e11,e12,e13,e14,e15)          \
-  _mm512_set_epi32((e15),(e14),(e13),(e12),(e11),(e10),(e9),(e8),(e7),(e6), \
-                   (e5),(e4),(e3),(e2),(e1),(e0))
-
-static __inline__ __m512i __DEFAULT_FN_ATTRS512
-_mm512_set_epi64 (long long __A, long long __B, long long __C,
-     long long __D, long long __E, long long __F,
-     long long __G, long long __H)
-{
-  return __extension__ (__m512i) (__v8di)
-  { __H, __G, __F, __E, __D, __C, __B, __A };
-}
-
-#define _mm512_setr_epi64(e0,e1,e2,e3,e4,e5,e6,e7)           \
-  _mm512_set_epi64((e7),(e6),(e5),(e4),(e3),(e2),(e1),(e0))
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_set_pd (double __A, double __B, double __C, double __D,
-        double __E, double __F, double __G, double __H)
-{
-  return __extension__ (__m512d)
-  { __H, __G, __F, __E, __D, __C, __B, __A };
-}
-
-#define _mm512_setr_pd(e0,e1,e2,e3,e4,e5,e6,e7)              \
-  _mm512_set_pd((e7),(e6),(e5),(e4),(e3),(e2),(e1),(e0))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_set_ps (float __A, float __B, float __C, float __D,
-        float __E, float __F, float __G, float __H,
-        float __I, float __J, float __K, float __L,
-        float __M, float __N, float __O, float __P)
-{
-  return __extension__ (__m512)
-  { __P, __O, __N, __M, __L, __K, __J, __I,
-    __H, __G, __F, __E, __D, __C, __B, __A };
-}
-
-#define _mm512_setr_ps(e0,e1,e2,e3,e4,e5,e6,e7,e8,e9,e10,e11,e12,e13,e14,e15) \
-  _mm512_set_ps((e15),(e14),(e13),(e12),(e11),(e10),(e9),(e8),(e7),(e6),(e5), \
-                (e4),(e3),(e2),(e1),(e0))
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_abs_ps(__m512 __A)
-{
-  return (__m512)_mm512_and_epi32(_mm512_set1_epi32(0x7FFFFFFF),(__m512i)__A) ;
-}
-
-static __inline__ __m512 __DEFAULT_FN_ATTRS512
-_mm512_mask_abs_ps(__m512 __W, __mmask16 __K, __m512 __A)
-{
-  return (__m512)_mm512_mask_and_epi32((__m512i)__W, __K, _mm512_set1_epi32(0x7FFFFFFF),(__m512i)__A) ;
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_abs_pd(__m512d __A)
-{
-  return (__m512d)_mm512_and_epi64(_mm512_set1_epi64(0x7FFFFFFFFFFFFFFF),(__v8di)__A) ;
-}
-
-static __inline__ __m512d __DEFAULT_FN_ATTRS512
-_mm512_mask_abs_pd(__m512d __W, __mmask8 __K, __m512d __A)
-{
-  return (__m512d)_mm512_mask_and_epi64((__v8di)__W, __K, _mm512_set1_epi64(0x7FFFFFFFFFFFFFFF),(__v8di)__A);
-}
-
-/* Vector-reduction arithmetic accepts vectors as inputs and produces scalars as
- * outputs. This class of vector operation forms the basis of many scientific
- * computations. In vector-reduction arithmetic, the evaluation off is
- * independent of the order of the input elements of V.
-
- * Used bisection method. At each step, we partition the vector with previous
- * step in half, and the operation is performed on its two halves.
- * This takes log2(n) steps where n is the number of elements in the vector.
- */
-
-#define _mm512_mask_reduce_operator(op) \
-  __v4du __t1 = (__v4du)_mm512_extracti64x4_epi64(__W, 0); \
-  __v4du __t2 = (__v4du)_mm512_extracti64x4_epi64(__W, 1); \
-  __m256i __t3 = (__m256i)(__t1 op __t2); \
-  __v2du __t4 = (__v2du)_mm256_extracti128_si256(__t3, 0); \
-  __v2du __t5 = (__v2du)_mm256_extracti128_si256(__t3, 1); \
-  __v2du __t6 = __t4 op __t5; \
-  __v2du __t7 = __builtin_shufflevector(__t6, __t6, 1, 0); \
-  __v2du __t8 = __t6 op __t7; \
-  return __t8[0]
-
-static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_add_epi64(__m512i __W) {
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_epi64(__m512i __W) {
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_and_epi64(__m512i __W) {
-  _mm512_mask_reduce_operator(&);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_or_epi64(__m512i __W) {
-  _mm512_mask_reduce_operator(|);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_add_epi64(__mmask8 __M, __m512i __W) {
-  __W = _mm512_maskz_mov_epi64(__M, __W);
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_mul_epi64(__mmask8 __M, __m512i __W) {
-  __W = _mm512_mask_mov_epi64(_mm512_set1_epi64(1), __M, __W);
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_and_epi64(__mmask8 __M, __m512i __W) {
-  __W = _mm512_mask_mov_epi64(_mm512_set1_epi64(~0ULL), __M, __W);
-  _mm512_mask_reduce_operator(&);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_or_epi64(__mmask8 __M, __m512i __W) {
-  __W = _mm512_maskz_mov_epi64(__M, __W);
-  _mm512_mask_reduce_operator(|);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m256d __t1 = _mm512_extractf64x4_pd(__W, 0); \
-  __m256d __t2 = _mm512_extractf64x4_pd(__W, 1); \
-  __m256d __t3 = __t1 op __t2; \
-  __m128d __t4 = _mm256_extractf128_pd(__t3, 0); \
-  __m128d __t5 = _mm256_extractf128_pd(__t3, 1); \
-  __m128d __t6 = __t4 op __t5; \
-  __m128d __t7 = __builtin_shufflevector(__t6, __t6, 1, 0); \
-  __m128d __t8 = __t6 op __t7; \
-  return __t8[0]
-
-static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_reduce_add_pd(__m512d __W) {
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_pd(__m512d __W) {
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_add_pd(__mmask8 __M, __m512d __W) {
-  __W = _mm512_maskz_mov_pd(__M, __W);
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_mul_pd(__mmask8 __M, __m512d __W) {
-  __W = _mm512_mask_mov_pd(_mm512_set1_pd(1.0), __M, __W);
-  _mm512_mask_reduce_operator(*);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __v8su __t1 = (__v8su)_mm512_extracti64x4_epi64(__W, 0); \
-  __v8su __t2 = (__v8su)_mm512_extracti64x4_epi64(__W, 1); \
-  __m256i __t3 = (__m256i)(__t1 op __t2); \
-  __v4su __t4 = (__v4su)_mm256_extracti128_si256(__t3, 0); \
-  __v4su __t5 = (__v4su)_mm256_extracti128_si256(__t3, 1); \
-  __v4su __t6 = __t4 op __t5; \
-  __v4su __t7 = __builtin_shufflevector(__t6, __t6, 2, 3, 0, 1); \
-  __v4su __t8 = __t6 op __t7; \
-  __v4su __t9 = __builtin_shufflevector(__t8, __t8, 1, 0, 3, 2); \
-  __v4su __t10 = __t8 op __t9; \
-  return __t10[0]
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_add_epi32(__m512i __W) {
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_mul_epi32(__m512i __W) {
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_and_epi32(__m512i __W) {
-  _mm512_mask_reduce_operator(&);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_or_epi32(__m512i __W) {
-  _mm512_mask_reduce_operator(|);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_add_epi32( __mmask16 __M, __m512i __W) {
-  __W = _mm512_maskz_mov_epi32(__M, __W);
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_mul_epi32( __mmask16 __M, __m512i __W) {
-  __W = _mm512_mask_mov_epi32(_mm512_set1_epi32(1), __M, __W);
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_and_epi32( __mmask16 __M, __m512i __W) {
-  __W = _mm512_mask_mov_epi32(_mm512_set1_epi32(~0U), __M, __W);
-  _mm512_mask_reduce_operator(&);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_or_epi32(__mmask16 __M, __m512i __W) {
-  __W = _mm512_maskz_mov_epi32(__M, __W);
-  _mm512_mask_reduce_operator(|);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m256 __t1 = (__m256)_mm512_extractf64x4_pd((__m512d)__W, 0); \
-  __m256 __t2 = (__m256)_mm512_extractf64x4_pd((__m512d)__W, 1); \
-  __m256 __t3 = __t1 op __t2; \
-  __m128 __t4 = _mm256_extractf128_ps(__t3, 0); \
-  __m128 __t5 = _mm256_extractf128_ps(__t3, 1); \
-  __m128 __t6 = __t4 op __t5; \
-  __m128 __t7 = __builtin_shufflevector(__t6, __t6, 2, 3, 0, 1); \
-  __m128 __t8 = __t6 op __t7; \
-  __m128 __t9 = __builtin_shufflevector(__t8, __t8, 1, 0, 3, 2); \
-  __m128 __t10 = __t8 op __t9; \
-  return __t10[0]
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_reduce_add_ps(__m512 __W) {
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_reduce_mul_ps(__m512 __W) {
-  _mm512_mask_reduce_operator(*);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_add_ps(__mmask16 __M, __m512 __W) {
-  __W = _mm512_maskz_mov_ps(__M, __W);
-  _mm512_mask_reduce_operator(+);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_mul_ps(__mmask16 __M, __m512 __W) {
-  __W = _mm512_mask_mov_ps(_mm512_set1_ps(1.0f), __M, __W);
-  _mm512_mask_reduce_operator(*);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m512i __t1 = (__m512i)__builtin_shufflevector((__v8di)__V, (__v8di)__V, 4, 5, 6, 7, 0, 1, 2, 3); \
-  __m512i __t2 = _mm512_##op(__V, __t1); \
-  __m512i __t3 = (__m512i)__builtin_shufflevector((__v8di)__t2, (__v8di)__t2, 2, 3, 0, 1, 6, 7, 4, 5); \
-  __m512i __t4 = _mm512_##op(__t2, __t3); \
-  __m512i __t5 = (__m512i)__builtin_shufflevector((__v8di)__t4, (__v8di)__t4, 1, 0, 3, 2, 5, 4, 7, 6); \
-  __v8di __t6 = (__v8di)_mm512_##op(__t4, __t5); \
-  return __t6[0]
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_epi64(__m512i __V) {
-  _mm512_mask_reduce_operator(max_epi64);
-}
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_epu64(__m512i __V) {
-  _mm512_mask_reduce_operator(max_epu64);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_epi64(__m512i __V) {
-  _mm512_mask_reduce_operator(min_epi64);
-}
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_epu64(__m512i __V) {
-  _mm512_mask_reduce_operator(min_epu64);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_epi64(__mmask8 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(-__LONG_LONG_MAX__ - 1LL), __M, __V);
-  _mm512_mask_reduce_operator(max_epi64);
-}
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_epu64(__mmask8 __M, __m512i __V) {
-  __V = _mm512_maskz_mov_epi64(__M, __V);
-  _mm512_mask_reduce_operator(max_epu64);
-}
-
-static __inline__ long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_epi64(__mmask8 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(__LONG_LONG_MAX__), __M, __V);
-  _mm512_mask_reduce_operator(min_epi64);
-}
-
-static __inline__ unsigned long long __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_epu64(__mmask8 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi64(_mm512_set1_epi64(~0ULL), __M, __V);
-  _mm512_mask_reduce_operator(min_epu64);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m256i __t1 = _mm512_extracti64x4_epi64(__V, 0); \
-  __m256i __t2 = _mm512_extracti64x4_epi64(__V, 1); \
-  __m256i __t3 = _mm256_##op(__t1, __t2); \
-  __m128i __t4 = _mm256_extracti128_si256(__t3, 0); \
-  __m128i __t5 = _mm256_extracti128_si256(__t3, 1); \
-  __m128i __t6 = _mm_##op(__t4, __t5); \
-  __m128i __t7 = (__m128i)__builtin_shufflevector((__v4si)__t6, (__v4si)__t6, 2, 3, 0, 1); \
-  __m128i __t8 = _mm_##op(__t6, __t7); \
-  __m128i __t9 = (__m128i)__builtin_shufflevector((__v4si)__t8, (__v4si)__t8, 1, 0, 3, 2); \
-  __v4si __t10 = (__v4si)_mm_##op(__t8, __t9); \
-  return __t10[0]
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_epi32(__m512i __V) {
-  _mm512_mask_reduce_operator(max_epi32);
-}
-
-static __inline__ unsigned int __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_epu32(__m512i __V) {
-  _mm512_mask_reduce_operator(max_epu32);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_epi32(__m512i __V) {
-  _mm512_mask_reduce_operator(min_epi32);
-}
-
-static __inline__ unsigned int __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_epu32(__m512i __V) {
-  _mm512_mask_reduce_operator(min_epu32);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_epi32(__mmask16 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(-__INT_MAX__ - 1), __M, __V);
-  _mm512_mask_reduce_operator(max_epi32);
-}
-
-static __inline__ unsigned int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_epu32(__mmask16 __M, __m512i __V) {
-  __V = _mm512_maskz_mov_epi32(__M, __V);
-  _mm512_mask_reduce_operator(max_epu32);
-}
-
-static __inline__ int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_epi32(__mmask16 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(__INT_MAX__), __M, __V);
-  _mm512_mask_reduce_operator(min_epi32);
-}
-
-static __inline__ unsigned int __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_epu32(__mmask16 __M, __m512i __V) {
-  __V = _mm512_mask_mov_epi32(_mm512_set1_epi32(~0U), __M, __V);
-  _mm512_mask_reduce_operator(min_epu32);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m256d __t1 = _mm512_extractf64x4_pd(__V, 0); \
-  __m256d __t2 = _mm512_extractf64x4_pd(__V, 1); \
-  __m256d __t3 = _mm256_##op(__t1, __t2); \
-  __m128d __t4 = _mm256_extractf128_pd(__t3, 0); \
-  __m128d __t5 = _mm256_extractf128_pd(__t3, 1); \
-  __m128d __t6 = _mm_##op(__t4, __t5); \
-  __m128d __t7 = __builtin_shufflevector(__t6, __t6, 1, 0); \
-  __m128d __t8 = _mm_##op(__t6, __t7); \
-  return __t8[0]
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_pd(__m512d __V) {
-  _mm512_mask_reduce_operator(max_pd);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_pd(__m512d __V) {
-  _mm512_mask_reduce_operator(min_pd);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_pd(__mmask8 __M, __m512d __V) {
-  __V = _mm512_mask_mov_pd(_mm512_set1_pd(-__builtin_inf()), __M, __V);
-  _mm512_mask_reduce_operator(max_pd);
-}
-
-static __inline__ double __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_pd(__mmask8 __M, __m512d __V) {
-  __V = _mm512_mask_mov_pd(_mm512_set1_pd(__builtin_inf()), __M, __V);
-  _mm512_mask_reduce_operator(min_pd);
-}
-#undef _mm512_mask_reduce_operator
-
-#define _mm512_mask_reduce_operator(op) \
-  __m256 __t1 = (__m256)_mm512_extractf64x4_pd((__m512d)__V, 0); \
-  __m256 __t2 = (__m256)_mm512_extractf64x4_pd((__m512d)__V, 1); \
-  __m256 __t3 = _mm256_##op(__t1, __t2); \
-  __m128 __t4 = _mm256_extractf128_ps(__t3, 0); \
-  __m128 __t5 = _mm256_extractf128_ps(__t3, 1); \
-  __m128 __t6 = _mm_##op(__t4, __t5); \
-  __m128 __t7 = __builtin_shufflevector(__t6, __t6, 2, 3, 0, 1); \
-  __m128 __t8 = _mm_##op(__t6, __t7); \
-  __m128 __t9 = __builtin_shufflevector(__t8, __t8, 1, 0, 3, 2); \
-  __m128 __t10 = _mm_##op(__t8, __t9); \
-  return __t10[0]
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_reduce_max_ps(__m512 __V) {
-  _mm512_mask_reduce_operator(max_ps);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_reduce_min_ps(__m512 __V) {
-  _mm512_mask_reduce_operator(min_ps);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_max_ps(__mmask16 __M, __m512 __V) {
-  __V = _mm512_mask_mov_ps(_mm512_set1_ps(-__builtin_inff()), __M, __V);
-  _mm512_mask_reduce_operator(max_ps);
-}
-
-static __inline__ float __DEFAULT_FN_ATTRS512
-_mm512_mask_reduce_min_ps(__mmask16 __M, __m512 __V) {
-  __V = _mm512_mask_mov_ps(_mm512_set1_ps(__builtin_inff()), __M, __V);
-  _mm512_mask_reduce_operator(min_ps);
-}
-#undef _mm512_mask_reduce_operator
-
-#undef __DEFAULT_FN_ATTRS512
-#undef __DEFAULT_FN_ATTRS128
-#undef __DEFAULT_FN_ATTRS
-
-#endif /* __AVX512FINTRIN_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/cpuid.h b/linux-x86/lib64/clang/9.0.7/include/cpuid.h
deleted file mode 100644
index 02ffac2..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/cpuid.h
+++ /dev/null
@@ -1,302 +0,0 @@
-/*===---- cpuid.h - X86 cpu model detection --------------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#if !(__x86_64__ || __i386__)
-#error this header is for x86 only
-#endif
-
-/* Responses identification request with %eax 0 */
-/* AMD:     "AuthenticAMD" */
-#define signature_AMD_ebx 0x68747541
-#define signature_AMD_edx 0x69746e65
-#define signature_AMD_ecx 0x444d4163
-/* CENTAUR: "CentaurHauls" */
-#define signature_CENTAUR_ebx 0x746e6543
-#define signature_CENTAUR_edx 0x48727561
-#define signature_CENTAUR_ecx 0x736c7561
-/* CYRIX:   "CyrixInstead" */
-#define signature_CYRIX_ebx 0x69727943
-#define signature_CYRIX_edx 0x736e4978
-#define signature_CYRIX_ecx 0x64616574
-/* INTEL:   "GenuineIntel" */
-#define signature_INTEL_ebx 0x756e6547
-#define signature_INTEL_edx 0x49656e69
-#define signature_INTEL_ecx 0x6c65746e
-/* TM1:     "TransmetaCPU" */
-#define signature_TM1_ebx 0x6e617254
-#define signature_TM1_edx 0x74656d73
-#define signature_TM1_ecx 0x55504361
-/* TM2:     "GenuineTMx86" */
-#define signature_TM2_ebx 0x756e6547
-#define signature_TM2_edx 0x54656e69
-#define signature_TM2_ecx 0x3638784d
-/* NSC:     "Geode by NSC" */
-#define signature_NSC_ebx 0x646f6547
-#define signature_NSC_edx 0x43534e20
-#define signature_NSC_ecx 0x79622065
-/* NEXGEN:  "NexGenDriven" */
-#define signature_NEXGEN_ebx 0x4778654e
-#define signature_NEXGEN_edx 0x72446e65
-#define signature_NEXGEN_ecx 0x6e657669
-/* RISE:    "RiseRiseRise" */
-#define signature_RISE_ebx 0x65736952
-#define signature_RISE_edx 0x65736952
-#define signature_RISE_ecx 0x65736952
-/* SIS:     "SiS SiS SiS " */
-#define signature_SIS_ebx 0x20536953
-#define signature_SIS_edx 0x20536953
-#define signature_SIS_ecx 0x20536953
-/* UMC:     "UMC UMC UMC " */
-#define signature_UMC_ebx 0x20434d55
-#define signature_UMC_edx 0x20434d55
-#define signature_UMC_ecx 0x20434d55
-/* VIA:     "VIA VIA VIA " */
-#define signature_VIA_ebx 0x20414956
-#define signature_VIA_edx 0x20414956
-#define signature_VIA_ecx 0x20414956
-/* VORTEX:  "Vortex86 SoC" */
-#define signature_VORTEX_ebx 0x74726f56
-#define signature_VORTEX_edx 0x36387865
-#define signature_VORTEX_ecx 0x436f5320
-
-/* Features in %ecx for leaf 1 */
-#define bit_SSE3        0x00000001
-#define bit_PCLMULQDQ   0x00000002
-#define bit_PCLMUL      bit_PCLMULQDQ   /* for gcc compat */
-#define bit_DTES64      0x00000004
-#define bit_MONITOR     0x00000008
-#define bit_DSCPL       0x00000010
-#define bit_VMX         0x00000020
-#define bit_SMX         0x00000040
-#define bit_EIST        0x00000080
-#define bit_TM2         0x00000100
-#define bit_SSSE3       0x00000200
-#define bit_CNXTID      0x00000400
-#define bit_FMA         0x00001000
-#define bit_CMPXCHG16B  0x00002000
-#define bit_xTPR        0x00004000
-#define bit_PDCM        0x00008000
-#define bit_PCID        0x00020000
-#define bit_DCA         0x00040000
-#define bit_SSE41       0x00080000
-#define bit_SSE4_1      bit_SSE41       /* for gcc compat */
-#define bit_SSE42       0x00100000
-#define bit_SSE4_2      bit_SSE42       /* for gcc compat */
-#define bit_x2APIC      0x00200000
-#define bit_MOVBE       0x00400000
-#define bit_POPCNT      0x00800000
-#define bit_TSCDeadline 0x01000000
-#define bit_AESNI       0x02000000
-#define bit_AES         bit_AESNI       /* for gcc compat */
-#define bit_XSAVE       0x04000000
-#define bit_OSXSAVE     0x08000000
-#define bit_AVX         0x10000000
-#define bit_F16C        0x20000000
-#define bit_RDRND       0x40000000
-
-/* Features in %edx for leaf 1 */
-#define bit_FPU         0x00000001
-#define bit_VME         0x00000002
-#define bit_DE          0x00000004
-#define bit_PSE         0x00000008
-#define bit_TSC         0x00000010
-#define bit_MSR         0x00000020
-#define bit_PAE         0x00000040
-#define bit_MCE         0x00000080
-#define bit_CX8         0x00000100
-#define bit_CMPXCHG8B   bit_CX8         /* for gcc compat */
-#define bit_APIC        0x00000200
-#define bit_SEP         0x00000800
-#define bit_MTRR        0x00001000
-#define bit_PGE         0x00002000
-#define bit_MCA         0x00004000
-#define bit_CMOV        0x00008000
-#define bit_PAT         0x00010000
-#define bit_PSE36       0x00020000
-#define bit_PSN         0x00040000
-#define bit_CLFSH       0x00080000
-#define bit_DS          0x00200000
-#define bit_ACPI        0x00400000
-#define bit_MMX         0x00800000
-#define bit_FXSR        0x01000000
-#define bit_FXSAVE      bit_FXSR        /* for gcc compat */
-#define bit_SSE         0x02000000
-#define bit_SSE2        0x04000000
-#define bit_SS          0x08000000
-#define bit_HTT         0x10000000
-#define bit_TM          0x20000000
-#define bit_PBE         0x80000000
-
-/* Features in %ebx for leaf 7 sub-leaf 0 */
-#define bit_FSGSBASE    0x00000001
-#define bit_SGX         0x00000004
-#define bit_BMI         0x00000008
-#define bit_HLE         0x00000010
-#define bit_AVX2        0x00000020
-#define bit_SMEP        0x00000080
-#define bit_BMI2        0x00000100
-#define bit_ENH_MOVSB   0x00000200
-#define bit_INVPCID     0x00000400
-#define bit_RTM         0x00000800
-#define bit_MPX         0x00004000
-#define bit_AVX512F     0x00010000
-#define bit_AVX512DQ    0x00020000
-#define bit_RDSEED      0x00040000
-#define bit_ADX         0x00080000
-#define bit_AVX512IFMA  0x00200000
-#define bit_CLFLUSHOPT  0x00800000
-#define bit_CLWB        0x01000000
-#define bit_AVX512PF    0x04000000
-#define bit_AVX512ER    0x08000000
-#define bit_AVX512CD    0x10000000
-#define bit_SHA         0x20000000
-#define bit_AVX512BW    0x40000000
-#define bit_AVX512VL    0x80000000
-
-/* Features in %ecx for leaf 7 sub-leaf 0 */
-#define bit_PREFTCHWT1       0x00000001
-#define bit_AVX512VBMI       0x00000002
-#define bit_PKU              0x00000004
-#define bit_OSPKE            0x00000010
-#define bit_WAITPKG          0x00000020
-#define bit_AVX512VBMI2      0x00000040
-#define bit_SHSTK            0x00000080
-#define bit_GFNI             0x00000100
-#define bit_VAES             0x00000200
-#define bit_VPCLMULQDQ       0x00000400
-#define bit_AVX512VNNI       0x00000800
-#define bit_AVX512BITALG     0x00001000
-#define bit_AVX512VPOPCNTDQ  0x00004000
-#define bit_RDPID            0x00400000
-#define bit_CLDEMOTE         0x02000000
-#define bit_MOVDIRI          0x08000000
-#define bit_MOVDIR64B        0x10000000
-#define bit_ENQCMD           0x20000000
-
-/* Features in %edx for leaf 7 sub-leaf 0 */
-#define bit_AVX5124VNNIW  0x00000004
-#define bit_AVX5124FMAPS  0x00000008
-#define bit_PCONFIG       0x00040000
-#define bit_IBT           0x00100000
-
-/* Features in %eax for leaf 7 sub-leaf 1 */
-#define bit_AVX512BF16    0x00000020
-
-/* Features in %eax for leaf 13 sub-leaf 1 */
-#define bit_XSAVEOPT    0x00000001
-#define bit_XSAVEC      0x00000002
-#define bit_XSAVES      0x00000008
-
-/* Features in %eax for leaf 0x14 sub-leaf 0 */
-#define bit_PTWRITE     0x00000010
-
-/* Features in %ecx for leaf 0x80000001 */
-#define bit_LAHF_LM     0x00000001
-#define bit_ABM         0x00000020
-#define bit_LZCNT       bit_ABM        /* for gcc compat */
-#define bit_SSE4a       0x00000040
-#define bit_PRFCHW      0x00000100
-#define bit_XOP         0x00000800
-#define bit_LWP         0x00008000
-#define bit_FMA4        0x00010000
-#define bit_TBM         0x00200000
-#define bit_MWAITX      0x20000000
-
-/* Features in %edx for leaf 0x80000001 */
-#define bit_MMXEXT      0x00400000
-#define bit_LM          0x20000000
-#define bit_3DNOWP      0x40000000
-#define bit_3DNOW       0x80000000
-
-/* Features in %ebx for leaf 0x80000008 */
-#define bit_CLZERO      0x00000001
-#define bit_WBNOINVD    0x00000200
-
-
-#if __i386__
-#define __cpuid(__leaf, __eax, __ebx, __ecx, __edx) \
-    __asm("cpuid" : "=a"(__eax), "=b" (__ebx), "=c"(__ecx), "=d"(__edx) \
-                  : "0"(__leaf))
-
-#define __cpuid_count(__leaf, __count, __eax, __ebx, __ecx, __edx) \
-    __asm("cpuid" : "=a"(__eax), "=b" (__ebx), "=c"(__ecx), "=d"(__edx) \
-                  : "0"(__leaf), "2"(__count))
-#else
-/* x86-64 uses %rbx as the base register, so preserve it. */
-#define __cpuid(__leaf, __eax, __ebx, __ecx, __edx) \
-    __asm("  xchgq  %%rbx,%q1\n" \
-          "  cpuid\n" \
-          "  xchgq  %%rbx,%q1" \
-        : "=a"(__eax), "=r" (__ebx), "=c"(__ecx), "=d"(__edx) \
-        : "0"(__leaf))
-
-#define __cpuid_count(__leaf, __count, __eax, __ebx, __ecx, __edx) \
-    __asm("  xchgq  %%rbx,%q1\n" \
-          "  cpuid\n" \
-          "  xchgq  %%rbx,%q1" \
-        : "=a"(__eax), "=r" (__ebx), "=c"(__ecx), "=d"(__edx) \
-        : "0"(__leaf), "2"(__count))
-#endif
-
-static __inline int __get_cpuid_max (unsigned int __leaf, unsigned int *__sig)
-{
-    unsigned int __eax, __ebx, __ecx, __edx;
-#if __i386__
-    int __cpuid_supported;
-
-    __asm("  pushfl\n"
-          "  popl   %%eax\n"
-          "  movl   %%eax,%%ecx\n"
-          "  xorl   $0x00200000,%%eax\n"
-          "  pushl  %%eax\n"
-          "  popfl\n"
-          "  pushfl\n"
-          "  popl   %%eax\n"
-          "  movl   $0,%0\n"
-          "  cmpl   %%eax,%%ecx\n"
-          "  je     1f\n"
-          "  movl   $1,%0\n"
-          "1:"
-        : "=r" (__cpuid_supported) : : "eax", "ecx");
-    if (!__cpuid_supported)
-        return 0;
-#endif
-
-    __cpuid(__leaf, __eax, __ebx, __ecx, __edx);
-    if (__sig)
-        *__sig = __ebx;
-    return __eax;
-}
-
-static __inline int __get_cpuid (unsigned int __leaf, unsigned int *__eax,
-                                 unsigned int *__ebx, unsigned int *__ecx,
-                                 unsigned int *__edx)
-{
-    unsigned int __max_leaf = __get_cpuid_max(__leaf & 0x80000000, 0);
-
-    if (__max_leaf == 0 || __max_leaf < __leaf)
-        return 0;
-
-    __cpuid(__leaf, *__eax, *__ebx, *__ecx, *__edx);
-    return 1;
-}
-
-static __inline int __get_cpuid_count (unsigned int __leaf,
-                                       unsigned int __subleaf,
-                                       unsigned int *__eax, unsigned int *__ebx,
-                                       unsigned int *__ecx, unsigned int *__edx)
-{
-    unsigned int __max_leaf = __get_cpuid_max(__leaf & 0x80000000, 0);
-
-    if (__max_leaf == 0 || __max_leaf < __leaf)
-        return 0;
-
-    __cpuid_count(__leaf, __subleaf, *__eax, *__ebx, *__ecx, *__edx);
-    return 1;
-}
diff --git a/linux-x86/lib64/clang/9.0.7/include/emmintrin.h b/linux-x86/lib64/clang/9.0.7/include/emmintrin.h
deleted file mode 100644
index 3d55f5f..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/emmintrin.h
+++ /dev/null
@@ -1,4981 +0,0 @@
-/*===---- emmintrin.h - SSE2 intrinsics ------------------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __EMMINTRIN_H
-#define __EMMINTRIN_H
-
-#include <xmmintrin.h>
-
-typedef double __m128d __attribute__((__vector_size__(16), __aligned__(16)));
-typedef long long __m128i __attribute__((__vector_size__(16), __aligned__(16)));
-
-typedef double __m128d_u __attribute__((__vector_size__(16), __aligned__(1)));
-typedef long long __m128i_u __attribute__((__vector_size__(16), __aligned__(1)));
-
-/* Type defines.  */
-typedef double __v2df __attribute__ ((__vector_size__ (16)));
-typedef long long __v2di __attribute__ ((__vector_size__ (16)));
-typedef short __v8hi __attribute__((__vector_size__(16)));
-typedef char __v16qi __attribute__((__vector_size__(16)));
-
-/* Unsigned types */
-typedef unsigned long long __v2du __attribute__ ((__vector_size__ (16)));
-typedef unsigned short __v8hu __attribute__((__vector_size__(16)));
-typedef unsigned char __v16qu __attribute__((__vector_size__(16)));
-
-/* We need an explicitly signed variant for char. Note that this shouldn't
- * appear in the interface though. */
-typedef signed char __v16qs __attribute__((__vector_size__(16)));
-
-/* Define the default attributes for the functions in this file. */
-#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("sse2"), __min_vector_width__(128)))
-#define __DEFAULT_FN_ATTRS_MMX __attribute__((__always_inline__, __nodebug__, __target__("mmx,sse2"), __min_vector_width__(64)))
-
-/// Adds lower double-precision values in both operands and returns the
-///    sum in the lower 64 bits of the result. The upper 64 bits of the result
-///    are copied from the upper double-precision value of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VADDSD / ADDSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    sum of the lower 64 bits of both operands. The upper 64 bits are copied
-///    from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_add_sd(__m128d __a, __m128d __b)
-{
-  __a[0] += __b[0];
-  return __a;
-}
-
-/// Adds two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VADDPD / ADDPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] containing the sums of both
-///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_add_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2df)__a + (__v2df)__b);
-}
-
-/// Subtracts the lower double-precision value of the second operand
-///    from the lower double-precision value of the first operand and returns
-///    the difference in the lower 64 bits of the result. The upper 64 bits of
-///    the result are copied from the upper double-precision value of the first
-///    operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSUBSD / SUBSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the minuend.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing the subtrahend.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    difference of the lower 64 bits of both operands. The upper 64 bits are
-///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_sub_sd(__m128d __a, __m128d __b)
-{
-  __a[0] -= __b[0];
-  return __a;
-}
-
-/// Subtracts two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSUBPD / SUBPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the minuend.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing the subtrahend.
-/// \returns A 128-bit vector of [2 x double] containing the differences between
-///    both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_sub_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2df)__a - (__v2df)__b);
-}
-
-/// Multiplies lower double-precision values in both operands and returns
-///    the product in the lower 64 bits of the result. The upper 64 bits of the
-///    result are copied from the upper double-precision value of the first
-///    operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMULSD / MULSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    product of the lower 64 bits of both operands. The upper 64 bits are
-///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_mul_sd(__m128d __a, __m128d __b)
-{
-  __a[0] *= __b[0];
-  return __a;
-}
-
-/// Multiplies two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMULPD / MULPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \returns A 128-bit vector of [2 x double] containing the products of both
-///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_mul_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2df)__a * (__v2df)__b);
-}
-
-/// Divides the lower double-precision value of the first operand by the
-///    lower double-precision value of the second operand and returns the
-///    quotient in the lower 64 bits of the result. The upper 64 bits of the
-///    result are copied from the upper double-precision value of the first
-///    operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VDIVSD / DIVSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the dividend.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing divisor.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    quotient of the lower 64 bits of both operands. The upper 64 bits are
-///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_div_sd(__m128d __a, __m128d __b)
-{
-  __a[0] /= __b[0];
-  return __a;
-}
-
-/// Performs an element-by-element division of two 128-bit vectors of
-///    [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VDIVPD / DIVPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the dividend.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing the divisor.
-/// \returns A 128-bit vector of [2 x double] containing the quotients of both
-///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_div_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2df)__a / (__v2df)__b);
-}
-
-/// Calculates the square root of the lower double-precision value of
-///    the second operand and returns it in the lower 64 bits of the result.
-///    The upper 64 bits of the result are copied from the upper
-///    double-precision value of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSQRTSD / SQRTSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    upper 64 bits of this operand are copied to the upper 64 bits of the
-///    result.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    square root is calculated using the lower 64 bits of this operand.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    square root of the lower 64 bits of operand \a __b, and whose upper 64
-///    bits are copied from the upper 64 bits of operand \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_sqrt_sd(__m128d __a, __m128d __b)
-{
-  __m128d __c = __builtin_ia32_sqrtsd((__v2df)__b);
-  return __extension__ (__m128d) { __c[0], __a[1] };
-}
-
-/// Calculates the square root of the each of two values stored in a
-///    128-bit vector of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSQRTPD / SQRTPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector of [2 x double] containing the square roots of the
-///    values in the operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_sqrt_pd(__m128d __a)
-{
-  return __builtin_ia32_sqrtpd((__v2df)__a);
-}
-
-/// Compares lower 64-bit double-precision values of both operands, and
-///    returns the lesser of the pair of values in the lower 64-bits of the
-///    result. The upper 64 bits of the result are copied from the upper
-///    double-precision value of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMINSD / MINSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    lower 64 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    lower 64 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    minimum value between both operands. The upper 64 bits are copied from
-///    the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_min_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_minsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Performs element-by-element comparison of the two 128-bit vectors of
-///    [2 x double] and returns the vector containing the lesser of each pair of
-///    values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMINPD / MINPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \returns A 128-bit vector of [2 x double] containing the minimum values
-///    between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_min_pd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_minpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares lower 64-bit double-precision values of both operands, and
-///    returns the greater of the pair of values in the lower 64-bits of the
-///    result. The upper 64 bits of the result are copied from the upper
-///    double-precision value of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMAXSD / MAXSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    lower 64 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands. The
-///    lower 64 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    maximum value between both operands. The upper 64 bits are copied from
-///    the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_max_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_maxsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Performs element-by-element comparison of the two 128-bit vectors of
-///    [2 x double] and returns the vector containing the greater of each pair
-///    of values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMAXPD / MAXPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the operands.
-/// \returns A 128-bit vector of [2 x double] containing the maximum values
-///    between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_max_pd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_maxpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPAND / PAND </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] containing the bitwise AND of the
-///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_and_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2du)__a & (__v2du)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit vectors of [2 x double], using
-///    the one's complement of the values contained in the first source operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPANDN / PANDN </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the left source operand. The
-///    one's complement of this value is used in the bitwise AND.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing the right source operand.
-/// \returns A 128-bit vector of [2 x double] containing the bitwise AND of the
-///    values in the second operand and the one's complement of the first
-///    operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_andnot_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)(~(__v2du)__a & (__v2du)__b);
-}
-
-/// Performs a bitwise OR of two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPOR / POR </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] containing the bitwise OR of the
-///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_or_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2du)__a | (__v2du)__b);
-}
-
-/// Performs a bitwise XOR of two 128-bit vectors of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPXOR / PXOR </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [2 x double] containing one of the source operands.
-/// \returns A 128-bit vector of [2 x double] containing the bitwise XOR of the
-///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_xor_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)((__v2du)__a ^ (__v2du)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] for equality. Each comparison yields 0x0
-///    for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPEQPD / CMPEQPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpeq_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpeqpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are less than those in the second operand. Each comparison
-///    yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTPD / CMPLTPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmplt_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpltpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are less than or equal to those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLEPD / CMPLEPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmple_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmplepd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are greater than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTPD / CMPLTPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpgt_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpltpd((__v2df)__b, (__v2df)__a);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are greater than or equal to those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLEPD / CMPLEPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpge_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmplepd((__v2df)__b, (__v2df)__a);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are ordered with respect to those in the second operand.
-///
-///    A pair of double-precision values are "ordered" with respect to each
-///    other if neither value is a NaN. Each comparison yields 0x0 for false,
-///    0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPORDPD / CMPORDPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpord_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpordpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are unordered with respect to those in the second operand.
-///
-///    A pair of double-precision values are "unordered" with respect to each
-///    other if one or both values are NaN. Each comparison yields 0x0 for
-///    false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPUNORDPD / CMPUNORDPD </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpunord_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpunordpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are unequal to those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNEQPD / CMPNEQPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpneq_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpneqpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are not less than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTPD / CMPNLTPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnlt_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnltpd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are not less than or equal to those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLEPD / CMPNLEPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnle_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnlepd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are not greater than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTPD / CMPNLTPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpngt_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnltpd((__v2df)__b, (__v2df)__a);
-}
-
-/// Compares each of the corresponding double-precision values of the
-///    128-bit vectors of [2 x double] to determine if the values in the first
-///    operand are not greater than or equal to those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLEPD / CMPNLEPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \param __b
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector containing the comparison results.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnge_pd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnlepd((__v2df)__b, (__v2df)__a);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] for equality.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPEQSD / CMPEQSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpeq_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpeqsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTSD / CMPLTSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmplt_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpltsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLESD / CMPLESD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmple_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmplesd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than the corresponding value
-///    in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTSD / CMPLTSD </c> instruction.
-///
-/// \param __a
-///     A 128-bit vector of [2 x double]. The lower double-precision value is
-///     compared to the lower double-precision value of \a __b.
-/// \param __b
-///     A 128-bit vector of [2 x double]. The lower double-precision value is
-///     compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///     results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpgt_sd(__m128d __a, __m128d __b)
-{
-  __m128d __c = __builtin_ia32_cmpltsd((__v2df)__b, (__v2df)__a);
-  return __extension__ (__m128d) { __c[0], __a[1] };
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLESD / CMPLESD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpge_sd(__m128d __a, __m128d __b)
-{
-  __m128d __c = __builtin_ia32_cmplesd((__v2df)__b, (__v2df)__a);
-  return __extension__ (__m128d) { __c[0], __a[1] };
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is "ordered" with respect to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true. A pair
-///    of double-precision values are "ordered" with respect to each other if
-///    neither value is a NaN.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPORDSD / CMPORDSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpord_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpordsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is "unordered" with respect to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true. A pair
-///    of double-precision values are "unordered" with respect to each other if
-///    one or both values are NaN.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPUNORDSD / CMPUNORDSD </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpunord_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpunordsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is unequal to the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNEQSD / CMPNEQSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpneq_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpneqsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is not less than the corresponding
-///    value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTSD / CMPNLTSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnlt_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnltsd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is not less than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLESD / CMPNLESD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns  A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnle_sd(__m128d __a, __m128d __b)
-{
-  return (__m128d)__builtin_ia32_cmpnlesd((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is not greater than the corresponding
-///    value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTSD / CMPNLTSD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpngt_sd(__m128d __a, __m128d __b)
-{
-  __m128d __c = __builtin_ia32_cmpnltsd((__v2df)__b, (__v2df)__a);
-  return __extension__ (__m128d) { __c[0], __a[1] };
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is not greater than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0x0 for false, 0xFFFFFFFFFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLESD / CMPNLESD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns A 128-bit vector. The lower 64 bits contains the comparison
-///    results. The upper 64 bits are copied from the upper 64 bits of \a __a.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cmpnge_sd(__m128d __a, __m128d __b)
-{
-  __m128d __c = __builtin_ia32_cmpnlesd((__v2df)__b, (__v2df)__a);
-  return __extension__ (__m128d) { __c[0], __a[1] };
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] for equality.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comieq_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdeq((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comilt_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdlt((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///     A 128-bit vector of [2 x double]. The lower double-precision value is
-///     compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comile_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdle((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than the corresponding value
-///    in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comigt_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdgt((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comige_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdge((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is unequal to the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two
-///    lower double-precision values is NaN, 1 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISD / COMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 1 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comineq_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_comisdneq((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] for equality. The
-///    comparison yields 0 for false, 1 for true.
-///
-///    If either of the two lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomieq_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdeq((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two lower
-///    double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomilt_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdlt((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is less than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two lower
-///    double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///     A 128-bit vector of [2 x double]. The lower double-precision value is
-///     compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomile_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdle((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than the corresponding value
-///    in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two lower
-///    double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///     A 128-bit vector of [2 x double]. The lower double-precision value is
-///     compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomigt_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdgt((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is greater than or equal to the
-///    corresponding value in the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true.  If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower double-precision values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomige_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdge((__v2df)__a, (__v2df)__b);
-}
-
-/// Compares the lower double-precision floating-point values in each of
-///    the two 128-bit floating-point vectors of [2 x double] to determine if
-///    the value in the first parameter is unequal to the corresponding value in
-///    the second parameter.
-///
-///    The comparison yields 0 for false, 1 for true. If either of the two lower
-///    double-precision values is NaN, 1 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISD / UCOMISD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __b.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision value is
-///    compared to the lower double-precision value of \a __a.
-/// \returns An integer containing the comparison result. If either of the two
-///    lower double-precision values is NaN, 1 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomineq_sd(__m128d __a, __m128d __b)
-{
-  return __builtin_ia32_ucomisdneq((__v2df)__a, (__v2df)__b);
-}
-
-/// Converts the two double-precision floating-point elements of a
-///    128-bit vector of [2 x double] into two single-precision floating-point
-///    values, returned in the lower 64 bits of a 128-bit vector of [4 x float].
-///    The upper 64 bits of the result vector are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTPD2PS / CVTPD2PS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
-///    converted values. The upper 64 bits are set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtpd_ps(__m128d __a)
-{
-  return __builtin_ia32_cvtpd2ps((__v2df)__a);
-}
-
-/// Converts the lower two single-precision floating-point elements of a
-///    128-bit vector of [4 x float] into two double-precision floating-point
-///    values, returned in a 128-bit vector of [2 x double]. The upper two
-///    elements of the input vector are unused.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTPS2PD / CVTPS2PD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower two single-precision
-///    floating-point elements are converted to double-precision values. The
-///    upper two elements are unused.
-/// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cvtps_pd(__m128 __a)
-{
-  return (__m128d) __builtin_convertvector(
-      __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 0, 1), __v2df);
-}
-
-/// Converts the lower two integer elements of a 128-bit vector of
-///    [4 x i32] into two double-precision floating-point values, returned in a
-///    128-bit vector of [2 x double].
-///
-///    The upper two elements of the input vector are unused.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTDQ2PD / CVTDQ2PD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector of [4 x i32]. The lower two integer elements are
-///    converted to double-precision values.
-///
-///    The upper two elements are unused.
-/// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cvtepi32_pd(__m128i __a)
-{
-  return (__m128d) __builtin_convertvector(
-      __builtin_shufflevector((__v4si)__a, (__v4si)__a, 0, 1), __v2df);
-}
-
-/// Converts the two double-precision floating-point elements of a
-///    128-bit vector of [2 x double] into two signed 32-bit integer values,
-///    returned in the lower 64 bits of a 128-bit vector of [4 x i32]. The upper
-///    64 bits of the result vector are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTPD2DQ / CVTPD2DQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector of [4 x i32] whose lower 64 bits contain the
-///    converted values. The upper 64 bits are set to zero.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvtpd_epi32(__m128d __a)
-{
-  return __builtin_ia32_cvtpd2dq((__v2df)__a);
-}
-
-/// Converts the low-order element of a 128-bit vector of [2 x double]
-///    into a 32-bit signed integer value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSD2SI / CVTSD2SI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower 64 bits are used in the
-///    conversion.
-/// \returns A 32-bit signed integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvtsd_si32(__m128d __a)
-{
-  return __builtin_ia32_cvtsd2si((__v2df)__a);
-}
-
-/// Converts the lower double-precision floating-point element of a
-///    128-bit vector of [2 x double], in the second parameter, into a
-///    single-precision floating-point value, returned in the lower 32 bits of a
-///    128-bit vector of [4 x float]. The upper 96 bits of the result vector are
-///    copied from the upper 96 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSD2SS / CVTSD2SS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The upper 96 bits of this parameter are
-///    copied to the upper 96 bits of the result.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower double-precision
-///    floating-point element is used in the conversion.
-/// \returns A 128-bit vector of [4 x float]. The lower 32 bits contain the
-///    converted value from the second parameter. The upper 96 bits are copied
-///    from the upper 96 bits of the first parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtsd_ss(__m128 __a, __m128d __b)
-{
-  return (__m128)__builtin_ia32_cvtsd2ss((__v4sf)__a, (__v2df)__b);
-}
-
-/// Converts a 32-bit signed integer value, in the second parameter, into
-///    a double-precision floating-point value, returned in the lower 64 bits of
-///    a 128-bit vector of [2 x double]. The upper 64 bits of the result vector
-///    are copied from the upper 64 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSI2SD / CVTSI2SD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The upper 64 bits of this parameter are
-///    copied to the upper 64 bits of the result.
-/// \param __b
-///    A 32-bit signed integer containing the value to be converted.
-/// \returns A 128-bit vector of [2 x double]. The lower 64 bits contain the
-///    converted value from the second parameter. The upper 64 bits are copied
-///    from the upper 64 bits of the first parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cvtsi32_sd(__m128d __a, int __b)
-{
-  __a[0] = __b;
-  return __a;
-}
-
-/// Converts the lower single-precision floating-point element of a
-///    128-bit vector of [4 x float], in the second parameter, into a
-///    double-precision floating-point value, returned in the lower 64 bits of
-///    a 128-bit vector of [2 x double]. The upper 64 bits of the result vector
-///    are copied from the upper 64 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSS2SD / CVTSS2SD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The upper 64 bits of this parameter are
-///    copied to the upper 64 bits of the result.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower single-precision
-///    floating-point element is used in the conversion.
-/// \returns A 128-bit vector of [2 x double]. The lower 64 bits contain the
-///    converted value from the second parameter. The upper 64 bits are copied
-///    from the upper 64 bits of the first parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cvtss_sd(__m128d __a, __m128 __b)
-{
-  __a[0] = __b[0];
-  return __a;
-}
-
-/// Converts the two double-precision floating-point elements of a
-///    128-bit vector of [2 x double] into two signed 32-bit integer values,
-///    returned in the lower 64 bits of a 128-bit vector of [4 x i32].
-///
-///    If the result of either conversion is inexact, the result is truncated
-///    (rounded towards zero) regardless of the current MXCSR setting. The upper
-///    64 bits of the result vector are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTPD2DQ / CVTTPD2DQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 128-bit vector of [4 x i32] whose lower 64 bits contain the
-///    converted values. The upper 64 bits are set to zero.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvttpd_epi32(__m128d __a)
-{
-  return (__m128i)__builtin_ia32_cvttpd2dq((__v2df)__a);
-}
-
-/// Converts the low-order element of a [2 x double] vector into a 32-bit
-///    signed integer value, truncating the result when it is inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTSD2SI / CVTTSD2SI </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower 64 bits are used in the
-///    conversion.
-/// \returns A 32-bit signed integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvttsd_si32(__m128d __a)
-{
-  return __builtin_ia32_cvttsd2si((__v2df)__a);
-}
-
-/// Converts the two double-precision floating-point elements of a
-///    128-bit vector of [2 x double] into two signed 32-bit integer values,
-///    returned in a 64-bit vector of [2 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPD2PI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 64-bit vector of [2 x i32] containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpd_pi32(__m128d __a)
-{
-  return (__m64)__builtin_ia32_cvtpd2pi((__v2df)__a);
-}
-
-/// Converts the two double-precision floating-point elements of a
-///    128-bit vector of [2 x double] into two signed 32-bit integer values,
-///    returned in a 64-bit vector of [2 x i32].
-///
-///    If the result of either conversion is inexact, the result is truncated
-///    (rounded towards zero) regardless of the current MXCSR setting.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTTPD2PI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double].
-/// \returns A 64-bit vector of [2 x i32] containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvttpd_pi32(__m128d __a)
-{
-  return (__m64)__builtin_ia32_cvttpd2pi((__v2df)__a);
-}
-
-/// Converts the two signed 32-bit integer elements of a 64-bit vector of
-///    [2 x i32] into two double-precision floating-point values, returned in a
-///    128-bit vector of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PD </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of [2 x i32].
-/// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpi32_pd(__m64 __a)
-{
-  return __builtin_ia32_cvtpi2pd((__v2si)__a);
-}
-
-/// Returns the low-order element of a 128-bit vector of [2 x double] as
-///    a double-precision floating-point value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower 64 bits are returned.
-/// \returns A double-precision floating-point value copied from the lower 64
-///    bits of \a __a.
-static __inline__ double __DEFAULT_FN_ATTRS
-_mm_cvtsd_f64(__m128d __a)
-{
-  return __a[0];
-}
-
-/// Loads a 128-bit floating-point vector of [2 x double] from an aligned
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPD / MOVAPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location has to be 16-byte aligned.
-/// \returns A 128-bit vector of [2 x double] containing the loaded values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_load_pd(double const *__dp)
-{
-  return *(__m128d*)__dp;
-}
-
-/// Loads a double-precision floating-point value from a specified memory
-///    location and duplicates it to both vector elements of a 128-bit vector of
-///    [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVDDUP / MOVDDUP </c> instruction.
-///
-/// \param __dp
-///    A pointer to a memory location containing a double-precision value.
-/// \returns A 128-bit vector of [2 x double] containing the loaded and
-///    duplicated values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_load1_pd(double const *__dp)
-{
-  struct __mm_load1_pd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  double __u = ((struct __mm_load1_pd_struct*)__dp)->__u;
-  return __extension__ (__m128d){ __u, __u };
-}
-
-#define        _mm_load_pd1(dp)        _mm_load1_pd(dp)
-
-/// Loads two double-precision values, in reverse order, from an aligned
-///    memory location into a 128-bit vector of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPD / MOVAPD </c> instruction +
-/// needed shuffling instructions. In AVX mode, the shuffling may be combined
-/// with the \c VMOVAPD, resulting in only a \c VPERMILPD instruction.
-///
-/// \param __dp
-///    A 16-byte aligned pointer to an array of double-precision values to be
-///    loaded in reverse order.
-/// \returns A 128-bit vector of [2 x double] containing the reversed loaded
-///    values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_loadr_pd(double const *__dp)
-{
-  __m128d __u = *(__m128d*)__dp;
-  return __builtin_shufflevector((__v2df)__u, (__v2df)__u, 1, 0);
-}
-
-/// Loads a 128-bit floating-point vector of [2 x double] from an
-///    unaligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVUPD / MOVUPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \returns A 128-bit vector of [2 x double] containing the loaded values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_loadu_pd(double const *__dp)
-{
-  struct __loadu_pd {
-    __m128d_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_pd*)__dp)->__v;
-}
-
-/// Loads a 64-bit integer value to the low element of a 128-bit integer
-///    vector and clears the upper element.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __a
-///    A pointer to a 64-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \returns A 128-bit vector of [2 x i64] containing the loaded value.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_loadu_si64(void const *__a)
-{
-  struct __loadu_si64 {
-    long long __v;
-  } __attribute__((__packed__, __may_alias__));
-  long long __u = ((struct __loadu_si64*)__a)->__v;
-  return __extension__ (__m128i)(__v2di){__u, 0LL};
-}
-
-/// Loads a 32-bit integer value to the low element of a 128-bit integer
-///    vector and clears the upper element.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
-///
-/// \param __a
-///    A pointer to a 32-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \returns A 128-bit vector of [4 x i32] containing the loaded value.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_loadu_si32(void const *__a)
-{
-  struct __loadu_si32 {
-    int __v;
-  } __attribute__((__packed__, __may_alias__));
-  int __u = ((struct __loadu_si32*)__a)->__v;
-  return __extension__ (__m128i)(__v4si){__u, 0, 0, 0};
-}
-
-/// Loads a 16-bit integer value to the low element of a 128-bit integer
-///    vector and clears the upper element.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic does not correspond to a specific instruction.
-///
-/// \param __a
-///    A pointer to a 16-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \returns A 128-bit vector of [8 x i16] containing the loaded value.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_loadu_si16(void const *__a)
-{
-  struct __loadu_si16 {
-    short __v;
-  } __attribute__((__packed__, __may_alias__));
-  short __u = ((struct __loadu_si16*)__a)->__v;
-  return __extension__ (__m128i)(__v8hi){__u, 0, 0, 0, 0, 0, 0, 0};
-}
-
-/// Loads a 64-bit double-precision value to the low element of a
-///    128-bit integer vector and clears the upper element.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVSD / MOVSD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a memory location containing a double-precision value.
-///    The address of the memory location does not have to be aligned.
-/// \returns A 128-bit vector of [2 x double] containing the loaded value.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_load_sd(double const *__dp)
-{
-  struct __mm_load_sd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  double __u = ((struct __mm_load_sd_struct*)__dp)->__u;
-  return __extension__ (__m128d){ __u, 0 };
-}
-
-/// Loads a double-precision value into the high-order bits of a 128-bit
-///    vector of [2 x double]. The low-order bits are copied from the low-order
-///    bits of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVHPD / MOVHPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [63:0] are written to bits [63:0] of the result.
-/// \param __dp
-///    A pointer to a 64-bit memory location containing a double-precision
-///    floating-point value that is loaded. The loaded value is written to bits
-///    [127:64] of the result. The address of the memory location does not have
-///    to be aligned.
-/// \returns A 128-bit vector of [2 x double] containing the moved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_loadh_pd(__m128d __a, double const *__dp)
-{
-  struct __mm_loadh_pd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  double __u = ((struct __mm_loadh_pd_struct*)__dp)->__u;
-  return __extension__ (__m128d){ __a[0], __u };
-}
-
-/// Loads a double-precision value into the low-order bits of a 128-bit
-///    vector of [2 x double]. The high-order bits are copied from the
-///    high-order bits of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVLPD / MOVLPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [127:64] are written to bits [127:64] of the result.
-/// \param __dp
-///    A pointer to a 64-bit memory location containing a double-precision
-///    floating-point value that is loaded. The loaded value is written to bits
-///    [63:0] of the result. The address of the memory location does not have to
-///    be aligned.
-/// \returns A 128-bit vector of [2 x double] containing the moved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_loadl_pd(__m128d __a, double const *__dp)
-{
-  struct __mm_loadl_pd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  double __u = ((struct __mm_loadl_pd_struct*)__dp)->__u;
-  return __extension__ (__m128d){ __u, __a[1] };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double] with
-///    unspecified content. This could be used as an argument to another
-///    intrinsic function where the argument is required but the value is not
-///    actually used.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \returns A 128-bit floating-point vector of [2 x double] with unspecified
-///    content.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_undefined_pd(void)
-{
-  return (__m128d)__builtin_ia32_undef128();
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double]. The lower
-///    64 bits of the vector are initialized with the specified double-precision
-///    floating-point value. The upper 64 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __w
-///    A double-precision floating-point value used to initialize the lower 64
-///    bits of the result.
-/// \returns An initialized 128-bit floating-point vector of [2 x double]. The
-///    lower 64 bits contain the value of the parameter. The upper 64 bits are
-///    set to zero.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_set_sd(double __w)
-{
-  return __extension__ (__m128d){ __w, 0 };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double], with each
-///    of the two double-precision floating-point vector elements set to the
-///    specified double-precision floating-point value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVDDUP / MOVLHPS </c> instruction.
-///
-/// \param __w
-///    A double-precision floating-point value used to initialize each vector
-///    element of the result.
-/// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_set1_pd(double __w)
-{
-  return __extension__ (__m128d){ __w, __w };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double], with each
-///    of the two double-precision floating-point vector elements set to the
-///    specified double-precision floating-point value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVDDUP / MOVLHPS </c> instruction.
-///
-/// \param __w
-///    A double-precision floating-point value used to initialize each vector
-///    element of the result.
-/// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_set_pd1(double __w)
-{
-  return _mm_set1_pd(__w);
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double]
-///    initialized with the specified double-precision floating-point values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKLPD / UNPCKLPD </c> instruction.
-///
-/// \param __w
-///    A double-precision floating-point value used to initialize the upper 64
-///    bits of the result.
-/// \param __x
-///    A double-precision floating-point value used to initialize the lower 64
-///    bits of the result.
-/// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_set_pd(double __w, double __x)
-{
-  return __extension__ (__m128d){ __x, __w };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double],
-///    initialized in reverse order with the specified double-precision
-///    floating-point values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKLPD / UNPCKLPD </c> instruction.
-///
-/// \param __w
-///    A double-precision floating-point value used to initialize the lower 64
-///    bits of the result.
-/// \param __x
-///    A double-precision floating-point value used to initialize the upper 64
-///    bits of the result.
-/// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_setr_pd(double __w, double __x)
-{
-  return __extension__ (__m128d){ __w, __x };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double]
-///    initialized to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instruction.
-///
-/// \returns An initialized 128-bit floating-point vector of [2 x double] with
-///    all elements set to zero.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_setzero_pd(void)
-{
-  return __extension__ (__m128d){ 0, 0 };
-}
-
-/// Constructs a 128-bit floating-point vector of [2 x double]. The lower
-///    64 bits are set to the lower 64 bits of the second parameter. The upper
-///    64 bits are set to the upper 64 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VBLENDPD / BLENDPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The upper 64 bits are written to the
-///    upper 64 bits of the result.
-/// \param __b
-///    A 128-bit vector of [2 x double]. The lower 64 bits are written to the
-///    lower 64 bits of the result.
-/// \returns A 128-bit vector of [2 x double] containing the moved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_move_sd(__m128d __a, __m128d __b)
-{
-  __a[0] = __b[0];
-  return __a;
-}
-
-/// Stores the lower 64 bits of a 128-bit vector of [2 x double] to a
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVSD / MOVSD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 64-bit memory location.
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_sd(double *__dp, __m128d __a)
-{
-  struct __mm_store_sd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __mm_store_sd_struct*)__dp)->__u = __a[0];
-}
-
-/// Moves packed double-precision values from a 128-bit vector of
-///    [2 x double] to a memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c>VMOVAPD / MOVAPS</c> instruction.
-///
-/// \param __dp
-///    A pointer to an aligned memory location that can store two
-///    double-precision values.
-/// \param __a
-///    A packed 128-bit vector of [2 x double] containing the values to be
-///    moved.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_pd(double *__dp, __m128d __a)
-{
-  *(__m128d*)__dp = __a;
-}
-
-/// Moves the lower 64 bits of a 128-bit vector of [2 x double] twice to
-///    the upper and lower 64 bits of a memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the
-///   <c> VMOVDDUP + VMOVAPD / MOVLHPS + MOVAPS </c> instruction.
-///
-/// \param __dp
-///    A pointer to a memory location that can store two double-precision
-///    values.
-/// \param __a
-///    A 128-bit vector of [2 x double] whose lower 64 bits are copied to each
-///    of the values in \a __dp.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store1_pd(double *__dp, __m128d __a)
-{
-  __a = __builtin_shufflevector((__v2df)__a, (__v2df)__a, 0, 0);
-  _mm_store_pd(__dp, __a);
-}
-
-/// Moves the lower 64 bits of a 128-bit vector of [2 x double] twice to
-///    the upper and lower 64 bits of a memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the
-///   <c> VMOVDDUP + VMOVAPD / MOVLHPS + MOVAPS </c> instruction.
-///
-/// \param __dp
-///    A pointer to a memory location that can store two double-precision
-///    values.
-/// \param __a
-///    A 128-bit vector of [2 x double] whose lower 64 bits are copied to each
-///    of the values in \a __dp.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_pd1(double *__dp, __m128d __a)
-{
-  _mm_store1_pd(__dp, __a);
-}
-
-/// Stores a 128-bit vector of [2 x double] into an unaligned memory
-///    location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVUPD / MOVUPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_pd(double *__dp, __m128d __a)
-{
-  struct __storeu_pd {
-    __m128d_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_pd*)__dp)->__v = __a;
-}
-
-/// Stores two double-precision values, in reverse order, from a 128-bit
-///    vector of [2 x double] to a 16-byte aligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to a shuffling instruction followed by a
-/// <c> VMOVAPD / MOVAPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 16-byte aligned memory location that can store two
-///    double-precision values.
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the values to be reversed and
-///    stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storer_pd(double *__dp, __m128d __a)
-{
-  __a = __builtin_shufflevector((__v2df)__a, (__v2df)__a, 1, 0);
-  *(__m128d *)__dp = __a;
-}
-
-/// Stores the upper 64 bits of a 128-bit vector of [2 x double] to a
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVHPD / MOVHPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 64-bit memory location.
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeh_pd(double *__dp, __m128d __a)
-{
-  struct __mm_storeh_pd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __mm_storeh_pd_struct*)__dp)->__u = __a[1];
-}
-
-/// Stores the lower 64 bits of a 128-bit vector of [2 x double] to a
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVLPD / MOVLPD </c> instruction.
-///
-/// \param __dp
-///    A pointer to a 64-bit memory location.
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storel_pd(double *__dp, __m128d __a)
-{
-  struct __mm_storeh_pd_struct {
-    double __u;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __mm_storeh_pd_struct*)__dp)->__u = __a[0];
-}
-
-/// Adds the corresponding elements of two 128-bit vectors of [16 x i8],
-///    saving the lower 8 bits of each sum in the corresponding element of a
-///    128-bit result vector of [16 x i8].
-///
-///    The integer elements of both parameters can be either signed or unsigned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDB / PADDB </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [16 x i8].
-/// \param __b
-///    A 128-bit vector of [16 x i8].
-/// \returns A 128-bit vector of [16 x i8] containing the sums of both
-///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_add_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v16qu)__a + (__v16qu)__b);
-}
-
-/// Adds the corresponding elements of two 128-bit vectors of [8 x i16],
-///    saving the lower 16 bits of each sum in the corresponding element of a
-///    128-bit result vector of [8 x i16].
-///
-///    The integer elements of both parameters can be either signed or unsigned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDW / PADDW </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [8 x i16].
-/// \param __b
-///    A 128-bit vector of [8 x i16].
-/// \returns A 128-bit vector of [8 x i16] containing the sums of both
-///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_add_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v8hu)__a + (__v8hu)__b);
-}
-
-/// Adds the corresponding elements of two 128-bit vectors of [4 x i32],
-///    saving the lower 32 bits of each sum in the corresponding element of a
-///    128-bit result vector of [4 x i32].
-///
-///    The integer elements of both parameters can be either signed or unsigned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDD / PADDD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x i32].
-/// \param __b
-///    A 128-bit vector of [4 x i32].
-/// \returns A 128-bit vector of [4 x i32] containing the sums of both
-///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_add_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v4su)__a + (__v4su)__b);
-}
-
-/// Adds two signed or unsigned 64-bit integer values, returning the
-///    lower 64 bits of the sum.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PADDQ </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer.
-/// \param __b
-///    A 64-bit integer.
-/// \returns A 64-bit integer containing the sum of both parameters.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_add_si64(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_paddq((__v1di)__a, (__v1di)__b);
-}
-
-/// Adds the corresponding elements of two 128-bit vectors of [2 x i64],
-///    saving the lower 64 bits of each sum in the corresponding element of a
-///    128-bit result vector of [2 x i64].
-///
-///    The integer elements of both parameters can be either signed or unsigned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDQ / PADDQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x i64].
-/// \param __b
-///    A 128-bit vector of [2 x i64].
-/// \returns A 128-bit vector of [2 x i64] containing the sums of both
-///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_add_epi64(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v2du)__a + (__v2du)__b);
-}
-
-/// Adds, with saturation, the corresponding elements of two 128-bit
-///    signed [16 x i8] vectors, saving each sum in the corresponding element of
-///    a 128-bit result vector of [16 x i8]. Positive sums greater than 0x7F are
-///    saturated to 0x7F. Negative sums less than 0x80 are saturated to 0x80.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDSB / PADDSB </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [16 x i8] vector.
-/// \param __b
-///    A 128-bit signed [16 x i8] vector.
-/// \returns A 128-bit signed [16 x i8] vector containing the saturated sums of
-///    both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_adds_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_paddsb128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Adds, with saturation, the corresponding elements of two 128-bit
-///    signed [8 x i16] vectors, saving each sum in the corresponding element of
-///    a 128-bit result vector of [8 x i16]. Positive sums greater than 0x7FFF
-///    are saturated to 0x7FFF. Negative sums less than 0x8000 are saturated to
-///    0x8000.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDSW / PADDSW </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [8 x i16] vector containing the saturated sums of
-///    both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_adds_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_paddsw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Adds, with saturation, the corresponding elements of two 128-bit
-///    unsigned [16 x i8] vectors, saving each sum in the corresponding element
-///    of a 128-bit result vector of [16 x i8]. Positive sums greater than 0xFF
-///    are saturated to 0xFF. Negative sums are saturated to 0x00.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDUSB / PADDUSB </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [16 x i8] vector.
-/// \param __b
-///    A 128-bit unsigned [16 x i8] vector.
-/// \returns A 128-bit unsigned [16 x i8] vector containing the saturated sums
-///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_adds_epu8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_paddusb128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Adds, with saturation, the corresponding elements of two 128-bit
-///    unsigned [8 x i16] vectors, saving each sum in the corresponding element
-///    of a 128-bit result vector of [8 x i16]. Positive sums greater than
-///    0xFFFF are saturated to 0xFFFF. Negative sums are saturated to 0x0000.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPADDUSB / PADDUSB </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [8 x i16] vector.
-/// \param __b
-///    A 128-bit unsigned [8 x i16] vector.
-/// \returns A 128-bit unsigned [8 x i16] vector containing the saturated sums
-///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_adds_epu16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_paddusw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Computes the rounded avarages of corresponding elements of two
-///    128-bit unsigned [16 x i8] vectors, saving each result in the
-///    corresponding element of a 128-bit result vector of [16 x i8].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPAVGB / PAVGB </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [16 x i8] vector.
-/// \param __b
-///    A 128-bit unsigned [16 x i8] vector.
-/// \returns A 128-bit unsigned [16 x i8] vector containing the rounded
-///    averages of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_avg_epu8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pavgb128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Computes the rounded avarages of corresponding elements of two
-///    128-bit unsigned [8 x i16] vectors, saving each result in the
-///    corresponding element of a 128-bit result vector of [8 x i16].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPAVGW / PAVGW </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [8 x i16] vector.
-/// \param __b
-///    A 128-bit unsigned [8 x i16] vector.
-/// \returns A 128-bit unsigned [8 x i16] vector containing the rounded
-///    averages of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_avg_epu16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pavgw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Multiplies the corresponding elements of two 128-bit signed [8 x i16]
-///    vectors, producing eight intermediate 32-bit signed integer products, and
-///    adds the consecutive pairs of 32-bit products to form a 128-bit signed
-///    [4 x i32] vector.
-///
-///    For example, bits [15:0] of both parameters are multiplied producing a
-///    32-bit product, bits [31:16] of both parameters are multiplied producing
-///    a 32-bit product, and the sum of those two products becomes bits [31:0]
-///    of the result.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMADDWD / PMADDWD </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [4 x i32] vector containing the sums of products
-///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_madd_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pmaddwd128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Compares corresponding elements of two 128-bit signed [8 x i16]
-///    vectors, saving the greater value from each comparison in the
-///    corresponding element of a 128-bit result vector of [8 x i16].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMAXSW / PMAXSW </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [8 x i16] vector containing the greater value of
-///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_max_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pmaxsw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Compares corresponding elements of two 128-bit unsigned [16 x i8]
-///    vectors, saving the greater value from each comparison in the
-///    corresponding element of a 128-bit result vector of [16 x i8].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMAXUB / PMAXUB </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [16 x i8] vector.
-/// \param __b
-///    A 128-bit unsigned [16 x i8] vector.
-/// \returns A 128-bit unsigned [16 x i8] vector containing the greater value of
-///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_max_epu8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pmaxub128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Compares corresponding elements of two 128-bit signed [8 x i16]
-///    vectors, saving the smaller value from each comparison in the
-///    corresponding element of a 128-bit result vector of [8 x i16].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMINSW / PMINSW </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [8 x i16] vector containing the smaller value of
-///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_min_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pminsw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Compares corresponding elements of two 128-bit unsigned [16 x i8]
-///    vectors, saving the smaller value from each comparison in the
-///    corresponding element of a 128-bit result vector of [16 x i8].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMINUB / PMINUB </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [16 x i8] vector.
-/// \param __b
-///    A 128-bit unsigned [16 x i8] vector.
-/// \returns A 128-bit unsigned [16 x i8] vector containing the smaller value of
-///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_min_epu8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pminub128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Multiplies the corresponding elements of two signed [8 x i16]
-///    vectors, saving the upper 16 bits of each 32-bit product in the
-///    corresponding element of a 128-bit signed [8 x i16] result vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMULHW / PMULHW </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [8 x i16] vector containing the upper 16 bits of
-///    each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_mulhi_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pmulhw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Multiplies the corresponding elements of two unsigned [8 x i16]
-///    vectors, saving the upper 16 bits of each 32-bit product in the
-///    corresponding element of a 128-bit unsigned [8 x i16] result vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMULHUW / PMULHUW </c> instruction.
-///
-/// \param __a
-///    A 128-bit unsigned [8 x i16] vector.
-/// \param __b
-///    A 128-bit unsigned [8 x i16] vector.
-/// \returns A 128-bit unsigned [8 x i16] vector containing the upper 16 bits
-///    of each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_mulhi_epu16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_pmulhuw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Multiplies the corresponding elements of two signed [8 x i16]
-///    vectors, saving the lower 16 bits of each 32-bit product in the
-///    corresponding element of a 128-bit signed [8 x i16] result vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMULLW / PMULLW </c> instruction.
-///
-/// \param __a
-///    A 128-bit signed [8 x i16] vector.
-/// \param __b
-///    A 128-bit signed [8 x i16] vector.
-/// \returns A 128-bit signed [8 x i16] vector containing the lower 16 bits of
-///    each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_mullo_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v8hu)__a * (__v8hu)__b);
-}
-
-/// Multiplies 32-bit unsigned integer values contained in the lower bits
-///    of the two 64-bit integer vectors and returns the 64-bit unsigned
-///    product.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMULUDQ </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer containing one of the source operands.
-/// \param __b
-///    A 64-bit integer containing one of the source operands.
-/// \returns A 64-bit integer vector containing the product of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_mul_su32(__m64 __a, __m64 __b)
-{
-  return __builtin_ia32_pmuludq((__v2si)__a, (__v2si)__b);
-}
-
-/// Multiplies 32-bit unsigned integer values contained in the lower
-///    bits of the corresponding elements of two [2 x i64] vectors, and returns
-///    the 64-bit products in the corresponding elements of a [2 x i64] vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMULUDQ / PMULUDQ </c> instruction.
-///
-/// \param __a
-///    A [2 x i64] vector containing one of the source operands.
-/// \param __b
-///    A [2 x i64] vector containing one of the source operands.
-/// \returns A [2 x i64] vector containing the product of both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_mul_epu32(__m128i __a, __m128i __b)
-{
-  return __builtin_ia32_pmuludq128((__v4si)__a, (__v4si)__b);
-}
-
-/// Computes the absolute differences of corresponding 8-bit integer
-///    values in two 128-bit vectors. Sums the first 8 absolute differences, and
-///    separately sums the second 8 absolute differences. Packs these two
-///    unsigned 16-bit integer sums into the upper and lower elements of a
-///    [2 x i64] vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSADBW / PSADBW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 128-bit integer vector containing one of the source operands.
-/// \returns A [2 x i64] vector containing the sums of the sets of absolute
-///    differences between both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sad_epu8(__m128i __a, __m128i __b)
-{
-  return __builtin_ia32_psadbw128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Subtracts the corresponding 8-bit integer values in the operands.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBB / PSUBB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sub_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v16qu)__a - (__v16qu)__b);
-}
-
-/// Subtracts the corresponding 16-bit integer values in the operands.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBW / PSUBW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sub_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v8hu)__a - (__v8hu)__b);
-}
-
-/// Subtracts the corresponding 32-bit integer values in the operands.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBD / PSUBD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sub_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v4su)__a - (__v4su)__b);
-}
-
-/// Subtracts signed or unsigned 64-bit integer values and writes the
-///    difference to the corresponding bits in the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PSUBQ </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing the minuend.
-/// \param __b
-///    A 64-bit integer vector containing the subtrahend.
-/// \returns A 64-bit integer vector containing the difference of the values in
-///    the operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_sub_si64(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_psubq((__v1di)__a, (__v1di)__b);
-}
-
-/// Subtracts the corresponding elements of two [2 x i64] vectors.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBQ / PSUBQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sub_epi64(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v2du)__a - (__v2du)__b);
-}
-
-/// Subtracts corresponding 8-bit signed integer values in the input and
-///    returns the differences in the corresponding bytes in the destination.
-///    Differences greater than 0x7F are saturated to 0x7F, and differences less
-///    than 0x80 are saturated to 0x80.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBSB / PSUBSB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_subs_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_psubsb128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Subtracts corresponding 16-bit signed integer values in the input and
-///    returns the differences in the corresponding bytes in the destination.
-///    Differences greater than 0x7FFF are saturated to 0x7FFF, and values less
-///    than 0x8000 are saturated to 0x8000.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBSW / PSUBSW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the differences of the values
-///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_subs_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_psubsw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Subtracts corresponding 8-bit unsigned integer values in the input
-///    and returns the differences in the corresponding bytes in the
-///    destination. Differences less than 0x00 are saturated to 0x00.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBUSB / PSUBUSB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the unsigned integer
-///    differences of the values in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_subs_epu8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_psubusb128((__v16qi)__a, (__v16qi)__b);
-}
-
-/// Subtracts corresponding 16-bit unsigned integer values in the input
-///    and returns the differences in the corresponding bytes in the
-///    destination. Differences less than 0x0000 are saturated to 0x0000.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSUBUSW / PSUBUSW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the minuends.
-/// \param __b
-///    A 128-bit integer vector containing the subtrahends.
-/// \returns A 128-bit integer vector containing the unsigned integer
-///    differences of the values in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_subs_epu16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_psubusw128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit integer vectors.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPAND / PAND </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 128-bit integer vector containing one of the source operands.
-/// \returns A 128-bit integer vector containing the bitwise AND of the values
-///    in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_and_si128(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v2du)__a & (__v2du)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit integer vectors, using the
-///    one's complement of the values contained in the first source operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPANDN / PANDN </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector containing the left source operand. The one's complement
-///    of this value is used in the bitwise AND.
-/// \param __b
-///    A 128-bit vector containing the right source operand.
-/// \returns A 128-bit integer vector containing the bitwise AND of the one's
-///    complement of the first operand and the values in the second operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_andnot_si128(__m128i __a, __m128i __b)
-{
-  return (__m128i)(~(__v2du)__a & (__v2du)__b);
-}
-/// Performs a bitwise OR of two 128-bit integer vectors.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPOR / POR </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 128-bit integer vector containing one of the source operands.
-/// \returns A 128-bit integer vector containing the bitwise OR of the values
-///    in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_or_si128(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v2du)__a | (__v2du)__b);
-}
-
-/// Performs a bitwise exclusive OR of two 128-bit integer vectors.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPXOR / PXOR </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 128-bit integer vector containing one of the source operands.
-/// \returns A 128-bit integer vector containing the bitwise exclusive OR of the
-///    values in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_xor_si128(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v2du)__a ^ (__v2du)__b);
-}
-
-/// Left-shifts the 128-bit integer vector operand by the specified
-///    number of bytes. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128i _mm_slli_si128(__m128i a, const int imm);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPSLLDQ / PSLLDQ </c> instruction.
-///
-/// \param a
-///    A 128-bit integer vector containing the source operand.
-/// \param imm
-///    An immediate value specifying the number of bytes to left-shift operand
-///    \a a.
-/// \returns A 128-bit integer vector containing the left-shifted value.
-#define _mm_slli_si128(a, imm) \
-  (__m128i)__builtin_ia32_pslldqi128_byteshift((__v2di)(__m128i)(a), (int)(imm))
-
-#define _mm_bslli_si128(a, imm) \
-  (__m128i)__builtin_ia32_pslldqi128_byteshift((__v2di)(__m128i)(a), (int)(imm))
-
-/// Left-shifts each 16-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLW / PSLLW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to left-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_slli_epi16(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_psllwi128((__v8hi)__a, __count);
-}
-
-/// Left-shifts each 16-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLW / PSLLW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to left-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sll_epi16(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_psllw128((__v8hi)__a, (__v8hi)__count);
-}
-
-/// Left-shifts each 32-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLD / PSLLD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to left-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_slli_epi32(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_pslldi128((__v4si)__a, __count);
-}
-
-/// Left-shifts each 32-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLD / PSLLD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to left-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sll_epi32(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_pslld128((__v4si)__a, (__v4si)__count);
-}
-
-/// Left-shifts each 64-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLQ / PSLLQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to left-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_slli_epi64(__m128i __a, int __count)
-{
-  return __builtin_ia32_psllqi128((__v2di)__a, __count);
-}
-
-/// Left-shifts each 64-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. Low-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSLLQ / PSLLQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to left-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sll_epi64(__m128i __a, __m128i __count)
-{
-  return __builtin_ia32_psllq128((__v2di)__a, (__v2di)__count);
-}
-
-/// Right-shifts each 16-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. High-order bits are filled with the sign
-///    bit of the initial value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRAW / PSRAW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to right-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srai_epi16(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_psrawi128((__v8hi)__a, __count);
-}
-
-/// Right-shifts each 16-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. High-order bits are filled with the sign
-///    bit of the initial value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRAW / PSRAW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to right-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sra_epi16(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_psraw128((__v8hi)__a, (__v8hi)__count);
-}
-
-/// Right-shifts each 32-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. High-order bits are filled with the sign
-///    bit of the initial value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRAD / PSRAD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to right-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srai_epi32(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_psradi128((__v4si)__a, __count);
-}
-
-/// Right-shifts each 32-bit value in the 128-bit integer vector operand
-///    by the specified number of bits. High-order bits are filled with the sign
-///    bit of the initial value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRAD / PSRAD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to right-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_sra_epi32(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_psrad128((__v4si)__a, (__v4si)__count);
-}
-
-/// Right-shifts the 128-bit integer vector operand by the specified
-///    number of bytes. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128i _mm_srli_si128(__m128i a, const int imm);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPSRLDQ / PSRLDQ </c> instruction.
-///
-/// \param a
-///    A 128-bit integer vector containing the source operand.
-/// \param imm
-///    An immediate value specifying the number of bytes to right-shift operand
-///    \a a.
-/// \returns A 128-bit integer vector containing the right-shifted value.
-#define _mm_srli_si128(a, imm) \
-  (__m128i)__builtin_ia32_psrldqi128_byteshift((__v2di)(__m128i)(a), (int)(imm))
-
-#define _mm_bsrli_si128(a, imm) \
-  (__m128i)__builtin_ia32_psrldqi128_byteshift((__v2di)(__m128i)(a), (int)(imm))
-
-/// Right-shifts each of 16-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLW / PSRLW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to right-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srli_epi16(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_psrlwi128((__v8hi)__a, __count);
-}
-
-/// Right-shifts each of 16-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLW / PSRLW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to right-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srl_epi16(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_psrlw128((__v8hi)__a, (__v8hi)__count);
-}
-
-/// Right-shifts each of 32-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLD / PSRLD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to right-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srli_epi32(__m128i __a, int __count)
-{
-  return (__m128i)__builtin_ia32_psrldi128((__v4si)__a, __count);
-}
-
-/// Right-shifts each of 32-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLD / PSRLD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to right-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srl_epi32(__m128i __a, __m128i __count)
-{
-  return (__m128i)__builtin_ia32_psrld128((__v4si)__a, (__v4si)__count);
-}
-
-/// Right-shifts each of 64-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLQ / PSRLQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    An integer value specifying the number of bits to right-shift each value
-///    in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srli_epi64(__m128i __a, int __count)
-{
-  return __builtin_ia32_psrlqi128((__v2di)__a, __count);
-}
-
-/// Right-shifts each of 64-bit values in the 128-bit integer vector
-///    operand by the specified number of bits. High-order bits are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPSRLQ / PSRLQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the source operand.
-/// \param __count
-///    A 128-bit integer vector in which bits [63:0] specify the number of bits
-///    to right-shift each value in operand \a __a.
-/// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_srl_epi64(__m128i __a, __m128i __count)
-{
-  return __builtin_ia32_psrlq128((__v2di)__a, (__v2di)__count);
-}
-
-/// Compares each of the corresponding 8-bit values of the 128-bit
-///    integer vectors for equality. Each comparison yields 0x0 for false, 0xFF
-///    for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPEQB / PCMPEQB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpeq_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v16qi)__a == (__v16qi)__b);
-}
-
-/// Compares each of the corresponding 16-bit values of the 128-bit
-///    integer vectors for equality. Each comparison yields 0x0 for false,
-///    0xFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPEQW / PCMPEQW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpeq_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v8hi)__a == (__v8hi)__b);
-}
-
-/// Compares each of the corresponding 32-bit values of the 128-bit
-///    integer vectors for equality. Each comparison yields 0x0 for false,
-///    0xFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPEQD / PCMPEQD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpeq_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v4si)__a == (__v4si)__b);
-}
-
-/// Compares each of the corresponding signed 8-bit values of the 128-bit
-///    integer vectors to determine if the values in the first operand are
-///    greater than those in the second operand. Each comparison yields 0x0 for
-///    false, 0xFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTB / PCMPGTB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpgt_epi8(__m128i __a, __m128i __b)
-{
-  /* This function always performs a signed comparison, but __v16qi is a char
-     which may be signed or unsigned, so use __v16qs. */
-  return (__m128i)((__v16qs)__a > (__v16qs)__b);
-}
-
-/// Compares each of the corresponding signed 16-bit values of the
-///    128-bit integer vectors to determine if the values in the first operand
-///    are greater than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTW / PCMPGTW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpgt_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v8hi)__a > (__v8hi)__b);
-}
-
-/// Compares each of the corresponding signed 32-bit values of the
-///    128-bit integer vectors to determine if the values in the first operand
-///    are greater than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTD / PCMPGTD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmpgt_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)((__v4si)__a > (__v4si)__b);
-}
-
-/// Compares each of the corresponding signed 8-bit values of the 128-bit
-///    integer vectors to determine if the values in the first operand are less
-///    than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTB / PCMPGTB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmplt_epi8(__m128i __a, __m128i __b)
-{
-  return _mm_cmpgt_epi8(__b, __a);
-}
-
-/// Compares each of the corresponding signed 16-bit values of the
-///    128-bit integer vectors to determine if the values in the first operand
-///    are less than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTW / PCMPGTW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmplt_epi16(__m128i __a, __m128i __b)
-{
-  return _mm_cmpgt_epi16(__b, __a);
-}
-
-/// Compares each of the corresponding signed 32-bit values of the
-///    128-bit integer vectors to determine if the values in the first operand
-///    are less than those in the second operand.
-///
-///    Each comparison yields 0x0 for false, 0xFFFFFFFF for true.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPCMPGTD / PCMPGTD </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __b
-///    A 128-bit integer vector.
-/// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cmplt_epi32(__m128i __a, __m128i __b)
-{
-  return _mm_cmpgt_epi32(__b, __a);
-}
-
-#ifdef __x86_64__
-/// Converts a 64-bit signed integer value from the second operand into a
-///    double-precision value and returns it in the lower element of a [2 x
-///    double] vector; the upper element of the returned vector is copied from
-///    the upper element of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSI2SD / CVTSI2SD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The upper 64 bits of this operand are
-///    copied to the upper 64 bits of the destination.
-/// \param __b
-///    A 64-bit signed integer operand containing the value to be converted.
-/// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
-///    converted value of the second operand. The upper 64 bits are copied from
-///    the upper 64 bits of the first operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_cvtsi64_sd(__m128d __a, long long __b)
-{
-  __a[0] = __b;
-  return __a;
-}
-
-/// Converts the first (lower) element of a vector of [2 x double] into a
-///    64-bit signed integer value, according to the current rounding mode.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSD2SI / CVTSD2SI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower 64 bits are used in the
-///    conversion.
-/// \returns A 64-bit signed integer containing the converted value.
-static __inline__ long long __DEFAULT_FN_ATTRS
-_mm_cvtsd_si64(__m128d __a)
-{
-  return __builtin_ia32_cvtsd2si64((__v2df)__a);
-}
-
-/// Converts the first (lower) element of a vector of [2 x double] into a
-///    64-bit signed integer value, truncating the result when it is inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTSD2SI / CVTTSD2SI </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. The lower 64 bits are used in the
-///    conversion.
-/// \returns A 64-bit signed integer containing the converted value.
-static __inline__ long long __DEFAULT_FN_ATTRS
-_mm_cvttsd_si64(__m128d __a)
-{
-  return __builtin_ia32_cvttsd2si64((__v2df)__a);
-}
-#endif
-
-/// Converts a vector of [4 x i32] into a vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTDQ2PS / CVTDQ2PS </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \returns A 128-bit vector of [4 x float] containing the converted values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtepi32_ps(__m128i __a)
-{
-  return (__m128)__builtin_convertvector((__v4si)__a, __v4sf);
-}
-
-/// Converts a vector of [4 x float] into a vector of [4 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTPS2DQ / CVTPS2DQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit integer vector of [4 x i32] containing the converted
-///    values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvtps_epi32(__m128 __a)
-{
-  return (__m128i)__builtin_ia32_cvtps2dq((__v4sf)__a);
-}
-
-/// Converts a vector of [4 x float] into a vector of [4 x i32],
-///    truncating the result when it is inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTPS2DQ / CVTTPS2DQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x i32] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvttps_epi32(__m128 __a)
-{
-  return (__m128i)__builtin_ia32_cvttps2dq((__v4sf)__a);
-}
-
-/// Returns a vector of [4 x i32] where the lowest element is the input
-///    operand and the remaining elements are zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
-///
-/// \param __a
-///    A 32-bit signed integer operand.
-/// \returns A 128-bit vector of [4 x i32].
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvtsi32_si128(int __a)
-{
-  return __extension__ (__m128i)(__v4si){ __a, 0, 0, 0 };
-}
-
-#ifdef __x86_64__
-/// Returns a vector of [2 x i64] where the lower element is the input
-///    operand and the upper element is zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __a
-///    A 64-bit signed integer operand containing the value to be converted.
-/// \returns A 128-bit vector of [2 x i64] containing the converted value.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_cvtsi64_si128(long long __a)
-{
-  return __extension__ (__m128i)(__v2di){ __a, 0 };
-}
-#endif
-
-/// Moves the least significant 32 bits of a vector of [4 x i32] to a
-///    32-bit signed integer value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
-///
-/// \param __a
-///    A vector of [4 x i32]. The least significant 32 bits are moved to the
-///    destination.
-/// \returns A 32-bit signed integer containing the moved value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvtsi128_si32(__m128i __a)
-{
-  __v4si __b = (__v4si)__a;
-  return __b[0];
-}
-
-#ifdef __x86_64__
-/// Moves the least significant 64 bits of a vector of [2 x i64] to a
-///    64-bit signed integer value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __a
-///    A vector of [2 x i64]. The least significant 64 bits are moved to the
-///    destination.
-/// \returns A 64-bit signed integer containing the moved value.
-static __inline__ long long __DEFAULT_FN_ATTRS
-_mm_cvtsi128_si64(__m128i __a)
-{
-  return __a[0];
-}
-#endif
-
-/// Moves packed integer values from an aligned 128-bit memory location
-///    to elements in a 128-bit integer vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVDQA / MOVDQA </c> instruction.
-///
-/// \param __p
-///    An aligned pointer to a memory location containing integer values.
-/// \returns A 128-bit integer vector containing the moved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_load_si128(__m128i const *__p)
-{
-  return *__p;
-}
-
-/// Moves packed integer values from an unaligned 128-bit memory location
-///    to elements in a 128-bit integer vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVDQU / MOVDQU </c> instruction.
-///
-/// \param __p
-///    A pointer to a memory location containing integer values.
-/// \returns A 128-bit integer vector containing the moved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_loadu_si128(__m128i_u const *__p)
-{
-  struct __loadu_si128 {
-    __m128i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_si128*)__p)->__v;
-}
-
-/// Returns a vector of [2 x i64] where the lower element is taken from
-///    the lower element of the operand, and the upper element is zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __p
-///    A 128-bit vector of [2 x i64]. Bits [63:0] are written to bits [63:0] of
-///    the destination.
-/// \returns A 128-bit vector of [2 x i64]. The lower order bits contain the
-///    moved value. The higher order bits are cleared.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_loadl_epi64(__m128i_u const *__p)
-{
-  struct __mm_loadl_epi64_struct {
-    long long __u;
-  } __attribute__((__packed__, __may_alias__));
-  return __extension__ (__m128i) { ((struct __mm_loadl_epi64_struct*)__p)->__u, 0};
-}
-
-/// Generates a 128-bit vector of [4 x i32] with unspecified content.
-///    This could be used as an argument to another intrinsic function where the
-///    argument is required but the value is not actually used.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \returns A 128-bit vector of [4 x i32] with unspecified content.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_undefined_si128(void)
-{
-  return (__m128i)__builtin_ia32_undef128();
-}
-
-/// Initializes both 64-bit values in a 128-bit vector of [2 x i64] with
-///    the specified 64-bit integer values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __q1
-///    A 64-bit integer value used to initialize the upper 64 bits of the
-///    destination vector of [2 x i64].
-/// \param __q0
-///    A 64-bit integer value used to initialize the lower 64 bits of the
-///    destination vector of [2 x i64].
-/// \returns An initialized 128-bit vector of [2 x i64] containing the values
-///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set_epi64x(long long __q1, long long __q0)
-{
-  return __extension__ (__m128i)(__v2di){ __q0, __q1 };
-}
-
-/// Initializes both 64-bit values in a 128-bit vector of [2 x i64] with
-///    the specified 64-bit integer values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __q1
-///    A 64-bit integer value used to initialize the upper 64 bits of the
-///    destination vector of [2 x i64].
-/// \param __q0
-///    A 64-bit integer value used to initialize the lower 64 bits of the
-///    destination vector of [2 x i64].
-/// \returns An initialized 128-bit vector of [2 x i64] containing the values
-///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set_epi64(__m64 __q1, __m64 __q0)
-{
-  return _mm_set_epi64x((long long)__q1, (long long)__q0);
-}
-
-/// Initializes the 32-bit values in a 128-bit vector of [4 x i32] with
-///    the specified 32-bit integer values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __i3
-///    A 32-bit integer value used to initialize bits [127:96] of the
-///    destination vector.
-/// \param __i2
-///    A 32-bit integer value used to initialize bits [95:64] of the destination
-///    vector.
-/// \param __i1
-///    A 32-bit integer value used to initialize bits [63:32] of the destination
-///    vector.
-/// \param __i0
-///    A 32-bit integer value used to initialize bits [31:0] of the destination
-///    vector.
-/// \returns An initialized 128-bit vector of [4 x i32] containing the values
-///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set_epi32(int __i3, int __i2, int __i1, int __i0)
-{
-  return __extension__ (__m128i)(__v4si){ __i0, __i1, __i2, __i3};
-}
-
-/// Initializes the 16-bit values in a 128-bit vector of [8 x i16] with
-///    the specified 16-bit integer values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __w7
-///    A 16-bit integer value used to initialize bits [127:112] of the
-///    destination vector.
-/// \param __w6
-///    A 16-bit integer value used to initialize bits [111:96] of the
-///    destination vector.
-/// \param __w5
-///    A 16-bit integer value used to initialize bits [95:80] of the destination
-///    vector.
-/// \param __w4
-///    A 16-bit integer value used to initialize bits [79:64] of the destination
-///    vector.
-/// \param __w3
-///    A 16-bit integer value used to initialize bits [63:48] of the destination
-///    vector.
-/// \param __w2
-///    A 16-bit integer value used to initialize bits [47:32] of the destination
-///    vector.
-/// \param __w1
-///    A 16-bit integer value used to initialize bits [31:16] of the destination
-///    vector.
-/// \param __w0
-///    A 16-bit integer value used to initialize bits [15:0] of the destination
-///    vector.
-/// \returns An initialized 128-bit vector of [8 x i16] containing the values
-///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set_epi16(short __w7, short __w6, short __w5, short __w4, short __w3, short __w2, short __w1, short __w0)
-{
-  return __extension__ (__m128i)(__v8hi){ __w0, __w1, __w2, __w3, __w4, __w5, __w6, __w7 };
-}
-
-/// Initializes the 8-bit values in a 128-bit vector of [16 x i8] with
-///    the specified 8-bit integer values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __b15
-///    Initializes bits [127:120] of the destination vector.
-/// \param __b14
-///    Initializes bits [119:112] of the destination vector.
-/// \param __b13
-///    Initializes bits [111:104] of the destination vector.
-/// \param __b12
-///    Initializes bits [103:96] of the destination vector.
-/// \param __b11
-///    Initializes bits [95:88] of the destination vector.
-/// \param __b10
-///    Initializes bits [87:80] of the destination vector.
-/// \param __b9
-///    Initializes bits [79:72] of the destination vector.
-/// \param __b8
-///    Initializes bits [71:64] of the destination vector.
-/// \param __b7
-///    Initializes bits [63:56] of the destination vector.
-/// \param __b6
-///    Initializes bits [55:48] of the destination vector.
-/// \param __b5
-///    Initializes bits [47:40] of the destination vector.
-/// \param __b4
-///    Initializes bits [39:32] of the destination vector.
-/// \param __b3
-///    Initializes bits [31:24] of the destination vector.
-/// \param __b2
-///    Initializes bits [23:16] of the destination vector.
-/// \param __b1
-///    Initializes bits [15:8] of the destination vector.
-/// \param __b0
-///    Initializes bits [7:0] of the destination vector.
-/// \returns An initialized 128-bit vector of [16 x i8] containing the values
-///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set_epi8(char __b15, char __b14, char __b13, char __b12, char __b11, char __b10, char __b9, char __b8, char __b7, char __b6, char __b5, char __b4, char __b3, char __b2, char __b1, char __b0)
-{
-  return __extension__ (__m128i)(__v16qi){ __b0, __b1, __b2, __b3, __b4, __b5, __b6, __b7, __b8, __b9, __b10, __b11, __b12, __b13, __b14, __b15 };
-}
-
-/// Initializes both values in a 128-bit integer vector with the
-///    specified 64-bit integer value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __q
-///    Integer value used to initialize the elements of the destination integer
-///    vector.
-/// \returns An initialized 128-bit integer vector of [2 x i64] with both
-///    elements containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set1_epi64x(long long __q)
-{
-  return _mm_set_epi64x(__q, __q);
-}
-
-/// Initializes both values in a 128-bit vector of [2 x i64] with the
-///    specified 64-bit value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __q
-///    A 64-bit value used to initialize the elements of the destination integer
-///    vector.
-/// \returns An initialized 128-bit vector of [2 x i64] with all elements
-///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set1_epi64(__m64 __q)
-{
-  return _mm_set_epi64(__q, __q);
-}
-
-/// Initializes all values in a 128-bit vector of [4 x i32] with the
-///    specified 32-bit value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __i
-///    A 32-bit value used to initialize the elements of the destination integer
-///    vector.
-/// \returns An initialized 128-bit vector of [4 x i32] with all elements
-///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set1_epi32(int __i)
-{
-  return _mm_set_epi32(__i, __i, __i, __i);
-}
-
-/// Initializes all values in a 128-bit vector of [8 x i16] with the
-///    specified 16-bit value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __w
-///    A 16-bit value used to initialize the elements of the destination integer
-///    vector.
-/// \returns An initialized 128-bit vector of [8 x i16] with all elements
-///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set1_epi16(short __w)
-{
-  return _mm_set_epi16(__w, __w, __w, __w, __w, __w, __w, __w);
-}
-
-/// Initializes all values in a 128-bit vector of [16 x i8] with the
-///    specified 8-bit value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __b
-///    An 8-bit value used to initialize the elements of the destination integer
-///    vector.
-/// \returns An initialized 128-bit vector of [16 x i8] with all elements
-///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_set1_epi8(char __b)
-{
-  return _mm_set_epi8(__b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b);
-}
-
-/// Constructs a 128-bit integer vector, initialized in reverse order
-///     with the specified 64-bit integral values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic does not correspond to a specific instruction.
-///
-/// \param __q0
-///    A 64-bit integral value used to initialize the lower 64 bits of the
-///    result.
-/// \param __q1
-///    A 64-bit integral value used to initialize the upper 64 bits of the
-///    result.
-/// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_setr_epi64(__m64 __q0, __m64 __q1)
-{
-  return _mm_set_epi64(__q1, __q0);
-}
-
-/// Constructs a 128-bit integer vector, initialized in reverse order
-///     with the specified 32-bit integral values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __i0
-///    A 32-bit integral value used to initialize bits [31:0] of the result.
-/// \param __i1
-///    A 32-bit integral value used to initialize bits [63:32] of the result.
-/// \param __i2
-///    A 32-bit integral value used to initialize bits [95:64] of the result.
-/// \param __i3
-///    A 32-bit integral value used to initialize bits [127:96] of the result.
-/// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_setr_epi32(int __i0, int __i1, int __i2, int __i3)
-{
-  return _mm_set_epi32(__i3, __i2, __i1, __i0);
-}
-
-/// Constructs a 128-bit integer vector, initialized in reverse order
-///     with the specified 16-bit integral values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __w0
-///    A 16-bit integral value used to initialize bits [15:0] of the result.
-/// \param __w1
-///    A 16-bit integral value used to initialize bits [31:16] of the result.
-/// \param __w2
-///    A 16-bit integral value used to initialize bits [47:32] of the result.
-/// \param __w3
-///    A 16-bit integral value used to initialize bits [63:48] of the result.
-/// \param __w4
-///    A 16-bit integral value used to initialize bits [79:64] of the result.
-/// \param __w5
-///    A 16-bit integral value used to initialize bits [95:80] of the result.
-/// \param __w6
-///    A 16-bit integral value used to initialize bits [111:96] of the result.
-/// \param __w7
-///    A 16-bit integral value used to initialize bits [127:112] of the result.
-/// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_setr_epi16(short __w0, short __w1, short __w2, short __w3, short __w4, short __w5, short __w6, short __w7)
-{
-  return _mm_set_epi16(__w7, __w6, __w5, __w4, __w3, __w2, __w1, __w0);
-}
-
-/// Constructs a 128-bit integer vector, initialized in reverse order
-///     with the specified 8-bit integral values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __b0
-///    An 8-bit integral value used to initialize bits [7:0] of the result.
-/// \param __b1
-///    An 8-bit integral value used to initialize bits [15:8] of the result.
-/// \param __b2
-///    An 8-bit integral value used to initialize bits [23:16] of the result.
-/// \param __b3
-///    An 8-bit integral value used to initialize bits [31:24] of the result.
-/// \param __b4
-///    An 8-bit integral value used to initialize bits [39:32] of the result.
-/// \param __b5
-///    An 8-bit integral value used to initialize bits [47:40] of the result.
-/// \param __b6
-///    An 8-bit integral value used to initialize bits [55:48] of the result.
-/// \param __b7
-///    An 8-bit integral value used to initialize bits [63:56] of the result.
-/// \param __b8
-///    An 8-bit integral value used to initialize bits [71:64] of the result.
-/// \param __b9
-///    An 8-bit integral value used to initialize bits [79:72] of the result.
-/// \param __b10
-///    An 8-bit integral value used to initialize bits [87:80] of the result.
-/// \param __b11
-///    An 8-bit integral value used to initialize bits [95:88] of the result.
-/// \param __b12
-///    An 8-bit integral value used to initialize bits [103:96] of the result.
-/// \param __b13
-///    An 8-bit integral value used to initialize bits [111:104] of the result.
-/// \param __b14
-///    An 8-bit integral value used to initialize bits [119:112] of the result.
-/// \param __b15
-///    An 8-bit integral value used to initialize bits [127:120] of the result.
-/// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_setr_epi8(char __b0, char __b1, char __b2, char __b3, char __b4, char __b5, char __b6, char __b7, char __b8, char __b9, char __b10, char __b11, char __b12, char __b13, char __b14, char __b15)
-{
-  return _mm_set_epi8(__b15, __b14, __b13, __b12, __b11, __b10, __b9, __b8, __b7, __b6, __b5, __b4, __b3, __b2, __b1, __b0);
-}
-
-/// Creates a 128-bit integer vector initialized to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instruction.
-///
-/// \returns An initialized 128-bit integer vector with all elements set to
-///    zero.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_setzero_si128(void)
-{
-  return __extension__ (__m128i)(__v2di){ 0LL, 0LL };
-}
-
-/// Stores a 128-bit integer vector to a memory location aligned on a
-///    128-bit boundary.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS </c> instruction.
-///
-/// \param __p
-///    A pointer to an aligned memory location that will receive the integer
-///    values.
-/// \param __b
-///    A 128-bit integer vector containing the values to be moved.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_si128(__m128i *__p, __m128i __b)
-{
-  *__p = __b;
-}
-
-/// Stores a 128-bit integer vector to an unaligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVUPS / MOVUPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a memory location that will receive the integer values.
-/// \param __b
-///    A 128-bit integer vector containing the values to be moved.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si128(__m128i_u *__p, __m128i __b)
-{
-  struct __storeu_si128 {
-    __m128i_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_si128*)__p)->__v = __b;
-}
-
-/// Stores a 64-bit integer value from the low element of a 128-bit integer
-///    vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __p
-///    A pointer to a 64-bit memory location. The address of the memory
-///    location does not have to be algned.
-/// \param __b
-///    A 128-bit integer vector containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si64(void const *__p, __m128i __b)
-{
-  struct __storeu_si64 {
-    long long __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_si64*)__p)->__v = ((__v2di)__b)[0];
-}
-
-/// Stores a 32-bit integer value from the low element of a 128-bit integer
-///    vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVD / MOVD </c> instruction.
-///
-/// \param __p
-///    A pointer to a 32-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \param __b
-///    A 128-bit integer vector containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si32(void const *__p, __m128i __b)
-{
-  struct __storeu_si32 {
-    int __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_si32*)__p)->__v = ((__v4si)__b)[0];
-}
-
-/// Stores a 16-bit integer value from the low element of a 128-bit integer
-///    vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic does not correspond to a specific instruction.
-///
-/// \param __p
-///    A pointer to a 16-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \param __b
-///    A 128-bit integer vector containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_si16(void const *__p, __m128i __b)
-{
-  struct __storeu_si16 {
-    short __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_si16*)__p)->__v = ((__v8hi)__b)[0];
-}
-
-/// Moves bytes selected by the mask from the first operand to the
-///    specified unaligned memory location. When a mask bit is 1, the
-///    corresponding byte is written, otherwise it is not written.
-///
-///    To minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon). Exception and trap behavior for elements not selected
-///    for storage to memory are implementation dependent.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMASKMOVDQU / MASKMOVDQU </c>
-///   instruction.
-///
-/// \param __d
-///    A 128-bit integer vector containing the values to be moved.
-/// \param __n
-///    A 128-bit integer vector containing the mask. The most significant bit of
-///    each byte represents the mask bits.
-/// \param __p
-///    A pointer to an unaligned 128-bit memory location where the specified
-///    values are moved.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_maskmoveu_si128(__m128i __d, __m128i __n, char *__p)
-{
-  __builtin_ia32_maskmovdqu((__v16qi)__d, (__v16qi)__n, __p);
-}
-
-/// Stores the lower 64 bits of a 128-bit integer vector of [2 x i64] to
-///    a memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVLPS / MOVLPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 64-bit memory location that will receive the lower 64 bits
-///    of the integer vector parameter.
-/// \param __a
-///    A 128-bit integer vector of [2 x i64]. The lower 64 bits contain the
-///    value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storel_epi64(__m128i_u *__p, __m128i __a)
-{
-  struct __mm_storel_epi64_struct {
-    long long __u;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __mm_storel_epi64_struct*)__p)->__u = __a[0];
-}
-
-/// Stores a 128-bit floating point vector of [2 x double] to a 128-bit
-///    aligned memory location.
-///
-///    To minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVNTPS / MOVNTPS </c> instruction.
-///
-/// \param __p
-///    A pointer to the 128-bit aligned memory location used to store the value.
-/// \param __a
-///    A vector of [2 x double] containing the 64-bit values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_stream_pd(double *__p, __m128d __a)
-{
-  __builtin_nontemporal_store((__v2df)__a, (__v2df*)__p);
-}
-
-/// Stores a 128-bit integer vector to a 128-bit aligned memory location.
-///
-///    To minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVNTPS / MOVNTPS </c> instruction.
-///
-/// \param __p
-///    A pointer to the 128-bit aligned memory location used to store the value.
-/// \param __a
-///    A 128-bit integer vector containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_stream_si128(__m128i *__p, __m128i __a)
-{
-  __builtin_nontemporal_store((__v2di)__a, (__v2di*)__p);
-}
-
-/// Stores a 32-bit integer value in the specified memory location.
-///
-///    To minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MOVNTI </c> instruction.
-///
-/// \param __p
-///    A pointer to the 32-bit memory location used to store the value.
-/// \param __a
-///    A 32-bit integer containing the value to be stored.
-static __inline__ void __attribute__((__always_inline__, __nodebug__, __target__("sse2")))
-_mm_stream_si32(int *__p, int __a)
-{
-  __builtin_ia32_movnti(__p, __a);
-}
-
-#ifdef __x86_64__
-/// Stores a 64-bit integer value in the specified memory location.
-///
-///    To minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MOVNTIQ </c> instruction.
-///
-/// \param __p
-///    A pointer to the 64-bit memory location used to store the value.
-/// \param __a
-///    A 64-bit integer containing the value to be stored.
-static __inline__ void __attribute__((__always_inline__, __nodebug__, __target__("sse2")))
-_mm_stream_si64(long long *__p, long long __a)
-{
-  __builtin_ia32_movnti64(__p, __a);
-}
-#endif
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-/// The cache line containing \a __p is flushed and invalidated from all
-///    caches in the coherency domain.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CLFLUSH </c> instruction.
-///
-/// \param __p
-///    A pointer to the memory location used to identify the cache line to be
-///    flushed.
-void _mm_clflush(void const * __p);
-
-/// Forces strong memory ordering (serialization) between load
-///    instructions preceding this instruction and load instructions following
-///    this instruction, ensuring the system completes all previous loads before
-///    executing subsequent loads.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> LFENCE </c> instruction.
-///
-void _mm_lfence(void);
-
-/// Forces strong memory ordering (serialization) between load and store
-///    instructions preceding this instruction and load and store instructions
-///    following this instruction, ensuring that the system completes all
-///    previous memory accesses before executing subsequent memory accesses.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MFENCE </c> instruction.
-///
-void _mm_mfence(void);
-
-#if defined(__cplusplus)
-} // extern "C"
-#endif
-
-/// Converts 16-bit signed integers from both 128-bit integer vector
-///    operands into 8-bit signed integers, and packs the results into the
-///    destination. Positive values greater than 0x7F are saturated to 0x7F.
-///    Negative values less than 0x80 are saturated to 0x80.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPACKSSWB / PACKSSWB </c> instruction.
-///
-/// \param __a
-///   A 128-bit integer vector of [8 x i16]. Each 16-bit element is treated as
-///   a signed integer and is converted to a 8-bit signed integer with
-///   saturation. Values greater than 0x7F are saturated to 0x7F. Values less
-///   than 0x80 are saturated to 0x80. The converted [8 x i8] values are
-///   written to the lower 64 bits of the result.
-/// \param __b
-///   A 128-bit integer vector of [8 x i16]. Each 16-bit element is treated as
-///   a signed integer and is converted to a 8-bit signed integer with
-///   saturation. Values greater than 0x7F are saturated to 0x7F. Values less
-///   than 0x80 are saturated to 0x80. The converted [8 x i8] values are
-///   written to the higher 64 bits of the result.
-/// \returns A 128-bit vector of [16 x i8] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_packs_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_packsswb128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Converts 32-bit signed integers from both 128-bit integer vector
-///    operands into 16-bit signed integers, and packs the results into the
-///    destination. Positive values greater than 0x7FFF are saturated to 0x7FFF.
-///    Negative values less than 0x8000 are saturated to 0x8000.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPACKSSDW / PACKSSDW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector of [4 x i32]. Each 32-bit element is treated as
-///    a signed integer and is converted to a 16-bit signed integer with
-///    saturation. Values greater than 0x7FFF are saturated to 0x7FFF. Values
-///    less than 0x8000 are saturated to 0x8000. The converted [4 x i16] values
-///    are written to the lower 64 bits of the result.
-/// \param __b
-///    A 128-bit integer vector of [4 x i32]. Each 32-bit element is treated as
-///    a signed integer and is converted to a 16-bit signed integer with
-///    saturation. Values greater than 0x7FFF are saturated to 0x7FFF. Values
-///    less than 0x8000 are saturated to 0x8000. The converted [4 x i16] values
-///    are written to the higher 64 bits of the result.
-/// \returns A 128-bit vector of [8 x i16] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_packs_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_packssdw128((__v4si)__a, (__v4si)__b);
-}
-
-/// Converts 16-bit signed integers from both 128-bit integer vector
-///    operands into 8-bit unsigned integers, and packs the results into the
-///    destination. Values greater than 0xFF are saturated to 0xFF. Values less
-///    than 0x00 are saturated to 0x00.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPACKUSWB / PACKUSWB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector of [8 x i16]. Each 16-bit element is treated as
-///    a signed integer and is converted to an 8-bit unsigned integer with
-///    saturation. Values greater than 0xFF are saturated to 0xFF. Values less
-///    than 0x00 are saturated to 0x00. The converted [8 x i8] values are
-///    written to the lower 64 bits of the result.
-/// \param __b
-///    A 128-bit integer vector of [8 x i16]. Each 16-bit element is treated as
-///    a signed integer and is converted to an 8-bit unsigned integer with
-///    saturation. Values greater than 0xFF are saturated to 0xFF. Values less
-///    than 0x00 are saturated to 0x00. The converted [8 x i8] values are
-///    written to the higher 64 bits of the result.
-/// \returns A 128-bit vector of [16 x i8] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_packus_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_ia32_packuswb128((__v8hi)__a, (__v8hi)__b);
-}
-
-/// Extracts 16 bits from a 128-bit integer vector of [8 x i16], using
-///    the immediate-value parameter as a selector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPEXTRW / PEXTRW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \param __imm
-///    An immediate value. Bits [2:0] selects values from \a __a to be assigned
-///    to bits[15:0] of the result. \n
-///    000: assign values from bits [15:0] of \a __a. \n
-///    001: assign values from bits [31:16] of \a __a. \n
-///    010: assign values from bits [47:32] of \a __a. \n
-///    011: assign values from bits [63:48] of \a __a. \n
-///    100: assign values from bits [79:64] of \a __a. \n
-///    101: assign values from bits [95:80] of \a __a. \n
-///    110: assign values from bits [111:96] of \a __a. \n
-///    111: assign values from bits [127:112] of \a __a.
-/// \returns An integer, whose lower 16 bits are selected from the 128-bit
-///    integer vector parameter and the remaining bits are assigned zeros.
-#define _mm_extract_epi16(a, imm) \
-  (int)(unsigned short)__builtin_ia32_vec_ext_v8hi((__v8hi)(__m128i)(a), \
-                                                   (int)(imm))
-
-/// Constructs a 128-bit integer vector by first making a copy of the
-///    128-bit integer vector parameter, and then inserting the lower 16 bits
-///    of an integer parameter into an offset specified by the immediate-value
-///    parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPINSRW / PINSRW </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector of [8 x i16]. This vector is copied to the
-///    result and then one of the eight elements in the result is replaced by
-///    the lower 16 bits of \a __b.
-/// \param __b
-///    An integer. The lower 16 bits of this parameter are written to the
-///    result beginning at an offset specified by \a __imm.
-/// \param __imm
-///    An immediate value specifying the bit offset in the result at which the
-///    lower 16 bits of \a __b are written.
-/// \returns A 128-bit integer vector containing the constructed values.
-#define _mm_insert_epi16(a, b, imm) \
-  (__m128i)__builtin_ia32_vec_set_v8hi((__v8hi)(__m128i)(a), (int)(b), \
-                                       (int)(imm))
-
-/// Copies the values of the most significant bits from each 8-bit
-///    element in a 128-bit integer vector of [16 x i8] to create a 16-bit mask
-///    value, zero-extends the value, and writes it to the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPMOVMSKB / PMOVMSKB </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector containing the values with bits to be extracted.
-/// \returns The most significant bits from each 8-bit element in \a __a,
-///    written to bits [15:0]. The other bits are assigned zeros.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_movemask_epi8(__m128i __a)
-{
-  return __builtin_ia32_pmovmskb128((__v16qi)__a);
-}
-
-/// Constructs a 128-bit integer vector by shuffling four 32-bit
-///    elements of a 128-bit integer vector parameter, using the immediate-value
-///    parameter as a specifier.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128i _mm_shuffle_epi32(__m128i a, const int imm);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPSHUFD / PSHUFD </c> instruction.
-///
-/// \param a
-///    A 128-bit integer vector containing the values to be copied.
-/// \param imm
-///    An immediate value containing an 8-bit value specifying which elements to
-///    copy from a. The destinations within the 128-bit destination are assigned
-///    values as follows: \n
-///    Bits [1:0] are used to assign values to bits [31:0] of the result. \n
-///    Bits [3:2] are used to assign values to bits [63:32] of the result. \n
-///    Bits [5:4] are used to assign values to bits [95:64] of the result. \n
-///    Bits [7:6] are used to assign values to bits [127:96] of the result. \n
-///    Bit value assignments: \n
-///    00: assign values from bits [31:0] of \a a. \n
-///    01: assign values from bits [63:32] of \a a. \n
-///    10: assign values from bits [95:64] of \a a. \n
-///    11: assign values from bits [127:96] of \a a.
-/// \returns A 128-bit integer vector containing the shuffled values.
-#define _mm_shuffle_epi32(a, imm) \
-  (__m128i)__builtin_ia32_pshufd((__v4si)(__m128i)(a), (int)(imm))
-
-/// Constructs a 128-bit integer vector by shuffling four lower 16-bit
-///    elements of a 128-bit integer vector of [8 x i16], using the immediate
-///    value parameter as a specifier.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128i _mm_shufflelo_epi16(__m128i a, const int imm);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPSHUFLW / PSHUFLW </c> instruction.
-///
-/// \param a
-///    A 128-bit integer vector of [8 x i16]. Bits [127:64] are copied to bits
-///    [127:64] of the result.
-/// \param imm
-///    An 8-bit immediate value specifying which elements to copy from \a a. \n
-///    Bits[1:0] are used to assign values to bits [15:0] of the result. \n
-///    Bits[3:2] are used to assign values to bits [31:16] of the result. \n
-///    Bits[5:4] are used to assign values to bits [47:32] of the result. \n
-///    Bits[7:6] are used to assign values to bits [63:48] of the result. \n
-///    Bit value assignments: \n
-///    00: assign values from bits [15:0] of \a a. \n
-///    01: assign values from bits [31:16] of \a a. \n
-///    10: assign values from bits [47:32] of \a a. \n
-///    11: assign values from bits [63:48] of \a a. \n
-/// \returns A 128-bit integer vector containing the shuffled values.
-#define _mm_shufflelo_epi16(a, imm) \
-  (__m128i)__builtin_ia32_pshuflw((__v8hi)(__m128i)(a), (int)(imm))
-
-/// Constructs a 128-bit integer vector by shuffling four upper 16-bit
-///    elements of a 128-bit integer vector of [8 x i16], using the immediate
-///    value parameter as a specifier.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128i _mm_shufflehi_epi16(__m128i a, const int imm);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPSHUFHW / PSHUFHW </c> instruction.
-///
-/// \param a
-///    A 128-bit integer vector of [8 x i16]. Bits [63:0] are copied to bits
-///    [63:0] of the result.
-/// \param imm
-///    An 8-bit immediate value specifying which elements to copy from \a a. \n
-///    Bits[1:0] are used to assign values to bits [79:64] of the result. \n
-///    Bits[3:2] are used to assign values to bits [95:80] of the result. \n
-///    Bits[5:4] are used to assign values to bits [111:96] of the result. \n
-///    Bits[7:6] are used to assign values to bits [127:112] of the result. \n
-///    Bit value assignments: \n
-///    00: assign values from bits [79:64] of \a a. \n
-///    01: assign values from bits [95:80] of \a a. \n
-///    10: assign values from bits [111:96] of \a a. \n
-///    11: assign values from bits [127:112] of \a a. \n
-/// \returns A 128-bit integer vector containing the shuffled values.
-#define _mm_shufflehi_epi16(a, imm) \
-  (__m128i)__builtin_ia32_pshufhw((__v8hi)(__m128i)(a), (int)(imm))
-
-/// Unpacks the high-order (index 8-15) values from two 128-bit vectors
-///    of [16 x i8] and interleaves them into a 128-bit vector of [16 x i8].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKHBW / PUNPCKHBW </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [16 x i8].
-///    Bits [71:64] are written to bits [7:0] of the result. \n
-///    Bits [79:72] are written to bits [23:16] of the result. \n
-///    Bits [87:80] are written to bits [39:32] of the result. \n
-///    Bits [95:88] are written to bits [55:48] of the result. \n
-///    Bits [103:96] are written to bits [71:64] of the result. \n
-///    Bits [111:104] are written to bits [87:80] of the result. \n
-///    Bits [119:112] are written to bits [103:96] of the result. \n
-///    Bits [127:120] are written to bits [119:112] of the result.
-/// \param __b
-///    A 128-bit vector of [16 x i8]. \n
-///    Bits [71:64] are written to bits [15:8] of the result. \n
-///    Bits [79:72] are written to bits [31:24] of the result. \n
-///    Bits [87:80] are written to bits [47:40] of the result. \n
-///    Bits [95:88] are written to bits [63:56] of the result. \n
-///    Bits [103:96] are written to bits [79:72] of the result. \n
-///    Bits [111:104] are written to bits [95:88] of the result. \n
-///    Bits [119:112] are written to bits [111:104] of the result. \n
-///    Bits [127:120] are written to bits [127:120] of the result.
-/// \returns A 128-bit vector of [16 x i8] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpackhi_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v16qi)__a, (__v16qi)__b, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
-}
-
-/// Unpacks the high-order (index 4-7) values from two 128-bit vectors of
-///    [8 x i16] and interleaves them into a 128-bit vector of [8 x i16].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKHWD / PUNPCKHWD </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [8 x i16].
-///    Bits [79:64] are written to bits [15:0] of the result. \n
-///    Bits [95:80] are written to bits [47:32] of the result. \n
-///    Bits [111:96] are written to bits [79:64] of the result. \n
-///    Bits [127:112] are written to bits [111:96] of the result.
-/// \param __b
-///    A 128-bit vector of [8 x i16].
-///    Bits [79:64] are written to bits [31:16] of the result. \n
-///    Bits [95:80] are written to bits [63:48] of the result. \n
-///    Bits [111:96] are written to bits [95:80] of the result. \n
-///    Bits [127:112] are written to bits [127:112] of the result.
-/// \returns A 128-bit vector of [8 x i16] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpackhi_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 4, 8+4, 5, 8+5, 6, 8+6, 7, 8+7);
-}
-
-/// Unpacks the high-order (index 2,3) values from two 128-bit vectors of
-///    [4 x i32] and interleaves them into a 128-bit vector of [4 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKHDQ / PUNPCKHDQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x i32]. \n
-///    Bits [95:64] are written to bits [31:0] of the destination. \n
-///    Bits [127:96] are written to bits [95:64] of the destination.
-/// \param __b
-///    A 128-bit vector of [4 x i32]. \n
-///    Bits [95:64] are written to bits [64:32] of the destination. \n
-///    Bits [127:96] are written to bits [127:96] of the destination.
-/// \returns A 128-bit vector of [4 x i32] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpackhi_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 2, 4+2, 3, 4+3);
-}
-
-/// Unpacks the high-order 64-bit elements from two 128-bit vectors of
-///    [2 x i64] and interleaves them into a 128-bit vector of [2 x i64].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKHQDQ / PUNPCKHQDQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x i64]. \n
-///    Bits [127:64] are written to bits [63:0] of the destination.
-/// \param __b
-///    A 128-bit vector of [2 x i64]. \n
-///    Bits [127:64] are written to bits [127:64] of the destination.
-/// \returns A 128-bit vector of [2 x i64] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpackhi_epi64(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v2di)__a, (__v2di)__b, 1, 2+1);
-}
-
-/// Unpacks the low-order (index 0-7) values from two 128-bit vectors of
-///    [16 x i8] and interleaves them into a 128-bit vector of [16 x i8].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKLBW / PUNPCKLBW </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [16 x i8]. \n
-///    Bits [7:0] are written to bits [7:0] of the result. \n
-///    Bits [15:8] are written to bits [23:16] of the result. \n
-///    Bits [23:16] are written to bits [39:32] of the result. \n
-///    Bits [31:24] are written to bits [55:48] of the result. \n
-///    Bits [39:32] are written to bits [71:64] of the result. \n
-///    Bits [47:40] are written to bits [87:80] of the result. \n
-///    Bits [55:48] are written to bits [103:96] of the result. \n
-///    Bits [63:56] are written to bits [119:112] of the result.
-/// \param __b
-///    A 128-bit vector of [16 x i8].
-///    Bits [7:0] are written to bits [15:8] of the result. \n
-///    Bits [15:8] are written to bits [31:24] of the result. \n
-///    Bits [23:16] are written to bits [47:40] of the result. \n
-///    Bits [31:24] are written to bits [63:56] of the result. \n
-///    Bits [39:32] are written to bits [79:72] of the result. \n
-///    Bits [47:40] are written to bits [95:88] of the result. \n
-///    Bits [55:48] are written to bits [111:104] of the result. \n
-///    Bits [63:56] are written to bits [127:120] of the result.
-/// \returns A 128-bit vector of [16 x i8] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpacklo_epi8(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v16qi)__a, (__v16qi)__b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7);
-}
-
-/// Unpacks the low-order (index 0-3) values from each of the two 128-bit
-///    vectors of [8 x i16] and interleaves them into a 128-bit vector of
-///    [8 x i16].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKLWD / PUNPCKLWD </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [8 x i16].
-///    Bits [15:0] are written to bits [15:0] of the result. \n
-///    Bits [31:16] are written to bits [47:32] of the result. \n
-///    Bits [47:32] are written to bits [79:64] of the result. \n
-///    Bits [63:48] are written to bits [111:96] of the result.
-/// \param __b
-///    A 128-bit vector of [8 x i16].
-///    Bits [15:0] are written to bits [31:16] of the result. \n
-///    Bits [31:16] are written to bits [63:48] of the result. \n
-///    Bits [47:32] are written to bits [95:80] of the result. \n
-///    Bits [63:48] are written to bits [127:112] of the result.
-/// \returns A 128-bit vector of [8 x i16] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpacklo_epi16(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 0, 8+0, 1, 8+1, 2, 8+2, 3, 8+3);
-}
-
-/// Unpacks the low-order (index 0,1) values from two 128-bit vectors of
-///    [4 x i32] and interleaves them into a 128-bit vector of [4 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKLDQ / PUNPCKLDQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x i32]. \n
-///    Bits [31:0] are written to bits [31:0] of the destination. \n
-///    Bits [63:32] are written to bits [95:64] of the destination.
-/// \param __b
-///    A 128-bit vector of [4 x i32]. \n
-///    Bits [31:0] are written to bits [64:32] of the destination. \n
-///    Bits [63:32] are written to bits [127:96] of the destination.
-/// \returns A 128-bit vector of [4 x i32] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpacklo_epi32(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 0, 4+0, 1, 4+1);
-}
-
-/// Unpacks the low-order 64-bit elements from two 128-bit vectors of
-///    [2 x i64] and interleaves them into a 128-bit vector of [2 x i64].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPUNPCKLQDQ / PUNPCKLQDQ </c>
-///   instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x i64]. \n
-///    Bits [63:0] are written to bits [63:0] of the destination. \n
-/// \param __b
-///    A 128-bit vector of [2 x i64]. \n
-///    Bits [63:0] are written to bits [127:64] of the destination. \n
-/// \returns A 128-bit vector of [2 x i64] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_unpacklo_epi64(__m128i __a, __m128i __b)
-{
-  return (__m128i)__builtin_shufflevector((__v2di)__a, (__v2di)__b, 0, 2+0);
-}
-
-/// Returns the lower 64 bits of a 128-bit integer vector as a 64-bit
-///    integer.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MOVDQ2Q </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector operand. The lower 64 bits are moved to the
-///    destination.
-/// \returns A 64-bit integer containing the lower 64 bits of the parameter.
-static __inline__ __m64 __DEFAULT_FN_ATTRS
-_mm_movepi64_pi64(__m128i __a)
-{
-  return (__m64)__a[0];
-}
-
-/// Moves the 64-bit operand to a 128-bit integer vector, zeroing the
-///    upper bits.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MOVD+VMOVQ </c> instruction.
-///
-/// \param __a
-///    A 64-bit value.
-/// \returns A 128-bit integer vector. The lower 64 bits contain the value from
-///    the operand. The upper 64 bits are assigned zeros.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_movpi64_epi64(__m64 __a)
-{
-  return __extension__ (__m128i)(__v2di){ (long long)__a, 0 };
-}
-
-/// Moves the lower 64 bits of a 128-bit integer vector to a 128-bit
-///    integer vector, zeroing the upper bits.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVQ / MOVQ </c> instruction.
-///
-/// \param __a
-///    A 128-bit integer vector operand. The lower 64 bits are moved to the
-///    destination.
-/// \returns A 128-bit integer vector. The lower 64 bits contain the value from
-///    the operand. The upper 64 bits are assigned zeros.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_move_epi64(__m128i __a)
-{
-  return __builtin_shufflevector((__v2di)__a, _mm_setzero_si128(), 0, 2);
-}
-
-/// Unpacks the high-order 64-bit elements from two 128-bit vectors of
-///    [2 x double] and interleaves them into a 128-bit vector of [2 x
-///    double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKHPD / UNPCKHPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [127:64] are written to bits [63:0] of the destination.
-/// \param __b
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [127:64] are written to bits [127:64] of the destination.
-/// \returns A 128-bit vector of [2 x double] containing the interleaved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_unpackhi_pd(__m128d __a, __m128d __b)
-{
-  return __builtin_shufflevector((__v2df)__a, (__v2df)__b, 1, 2+1);
-}
-
-/// Unpacks the low-order 64-bit elements from two 128-bit vectors
-///    of [2 x double] and interleaves them into a 128-bit vector of [2 x
-///    double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKLPD / UNPCKLPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [63:0] are written to bits [63:0] of the destination.
-/// \param __b
-///    A 128-bit vector of [2 x double]. \n
-///    Bits [63:0] are written to bits [127:64] of the destination.
-/// \returns A 128-bit vector of [2 x double] containing the interleaved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_unpacklo_pd(__m128d __a, __m128d __b)
-{
-  return __builtin_shufflevector((__v2df)__a, (__v2df)__b, 0, 2+0);
-}
-
-/// Extracts the sign bits of the double-precision values in the 128-bit
-///    vector of [2 x double], zero-extends the value, and writes it to the
-///    low-order bits of the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVMSKPD / MOVMSKPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [2 x double] containing the values with sign bits to
-///    be extracted.
-/// \returns The sign bits from each of the double-precision elements in \a __a,
-///    written to bits [1:0]. The remaining bits are assigned values of zero.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_movemask_pd(__m128d __a)
-{
-  return __builtin_ia32_movmskpd((__v2df)__a);
-}
-
-
-/// Constructs a 128-bit floating-point vector of [2 x double] from two
-///    128-bit vector parameters of [2 x double], using the immediate-value
-///     parameter as a specifier.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128d _mm_shuffle_pd(__m128d a, __m128d b, const int i);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VSHUFPD / SHUFPD </c> instruction.
-///
-/// \param a
-///    A 128-bit vector of [2 x double].
-/// \param b
-///    A 128-bit vector of [2 x double].
-/// \param i
-///    An 8-bit immediate value. The least significant two bits specify which
-///    elements to copy from \a a and \a b: \n
-///    Bit[0] = 0: lower element of \a a copied to lower element of result. \n
-///    Bit[0] = 1: upper element of \a a copied to lower element of result. \n
-///    Bit[1] = 0: lower element of \a b copied to upper element of result. \n
-///    Bit[1] = 1: upper element of \a b copied to upper element of result. \n
-/// \returns A 128-bit vector of [2 x double] containing the shuffled values.
-#define _mm_shuffle_pd(a, b, i) \
-  (__m128d)__builtin_ia32_shufpd((__v2df)(__m128d)(a), (__v2df)(__m128d)(b), \
-                                 (int)(i))
-
-/// Casts a 128-bit floating-point vector of [2 x double] into a 128-bit
-///    floating-point vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [2 x double].
-/// \returns A 128-bit floating-point vector of [4 x float] containing the same
-///    bitwise pattern as the parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_castpd_ps(__m128d __a)
-{
-  return (__m128)__a;
-}
-
-/// Casts a 128-bit floating-point vector of [2 x double] into a 128-bit
-///    integer vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [2 x double].
-/// \returns A 128-bit integer vector containing the same bitwise pattern as the
-///    parameter.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_castpd_si128(__m128d __a)
-{
-  return (__m128i)__a;
-}
-
-/// Casts a 128-bit floating-point vector of [4 x float] into a 128-bit
-///    floating-point vector of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float].
-/// \returns A 128-bit floating-point vector of [2 x double] containing the same
-///    bitwise pattern as the parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_castps_pd(__m128 __a)
-{
-  return (__m128d)__a;
-}
-
-/// Casts a 128-bit floating-point vector of [4 x float] into a 128-bit
-///    integer vector.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float].
-/// \returns A 128-bit integer vector containing the same bitwise pattern as the
-///    parameter.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
-_mm_castps_si128(__m128 __a)
-{
-  return (__m128i)__a;
-}
-
-/// Casts a 128-bit integer vector into a 128-bit floating-point vector
-///    of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \returns A 128-bit floating-point vector of [4 x float] containing the same
-///    bitwise pattern as the parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_castsi128_ps(__m128i __a)
-{
-  return (__m128)__a;
-}
-
-/// Casts a 128-bit integer vector into a 128-bit floating-point vector
-///    of [2 x double].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit integer vector.
-/// \returns A 128-bit floating-point vector of [2 x double] containing the same
-///    bitwise pattern as the parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS
-_mm_castsi128_pd(__m128i __a)
-{
-  return (__m128d)__a;
-}
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-/// Indicates that a spin loop is being executed for the purposes of
-///    optimizing power consumption during the loop.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PAUSE </c> instruction.
-///
-void _mm_pause(void);
-
-#if defined(__cplusplus)
-} // extern "C"
-#endif
-#undef __DEFAULT_FN_ATTRS
-#undef __DEFAULT_FN_ATTRS_MMX
-
-#define _MM_SHUFFLE2(x, y) (((x) << 1) | (y))
-
-#define _MM_DENORMALS_ZERO_ON   (0x0040)
-#define _MM_DENORMALS_ZERO_OFF  (0x0000)
-
-#define _MM_DENORMALS_ZERO_MASK (0x0040)
-
-#define _MM_GET_DENORMALS_ZERO_MODE() (_mm_getcsr() & _MM_DENORMALS_ZERO_MASK)
-#define _MM_SET_DENORMALS_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_DENORMALS_ZERO_MASK) | (x)))
-
-#endif /* __EMMINTRIN_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/omp.h b/linux-x86/lib64/clang/9.0.7/include/omp.h
deleted file mode 100644
index 7ba1859..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/omp.h
+++ /dev/null
@@ -1,370 +0,0 @@
-/*
- * include/50/omp.h.var
- */
-
-
-//===----------------------------------------------------------------------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-
-#ifndef __OMP_H
-#   define __OMP_H
-
-#   define KMP_VERSION_MAJOR    5
-#   define KMP_VERSION_MINOR    0
-#   define KMP_VERSION_BUILD    20140926
-#   define KMP_BUILD_DATE       "No_Timestamp"
-
-#   ifdef __cplusplus
-    extern "C" {
-#   endif
-
-#   define omp_set_affinity_format   ompc_set_affinity_format
-#   define omp_get_affinity_format   ompc_get_affinity_format
-#   define omp_display_affinity      ompc_display_affinity
-#   define omp_capture_affinity      ompc_capture_affinity
-
-#   if defined(_WIN32)
-#       define __KAI_KMPC_CONVENTION __cdecl
-#       ifndef __KMP_IMP
-#           define __KMP_IMP __declspec(dllimport)
-#       endif
-#   else
-#       define __KAI_KMPC_CONVENTION
-#       ifndef __KMP_IMP
-#           define __KMP_IMP
-#       endif
-#   endif
-
-    /* schedule kind constants */
-    typedef enum omp_sched_t {
-        omp_sched_static  = 1,
-        omp_sched_dynamic = 2,
-        omp_sched_guided  = 3,
-        omp_sched_auto    = 4,
-        omp_sched_monotonic = 0x80000000
-    } omp_sched_t;
-
-    /* set API functions */
-    extern void   __KAI_KMPC_CONVENTION  omp_set_num_threads (int);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_dynamic     (int);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_nested      (int);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_max_active_levels (int);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_schedule          (omp_sched_t, int);
-
-    /* query API functions */
-    extern int    __KAI_KMPC_CONVENTION  omp_get_num_threads  (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_dynamic      (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_nested       (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_max_threads  (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_thread_num   (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_num_procs    (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_in_parallel      (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_in_final         (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_active_level        (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_level               (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_ancestor_thread_num (int);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_team_size           (int);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_thread_limit        (void);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_max_active_levels   (void);
-    extern void   __KAI_KMPC_CONVENTION  omp_get_schedule            (omp_sched_t *, int *);
-    extern int    __KAI_KMPC_CONVENTION  omp_get_max_task_priority   (void);
-
-    /* lock API functions */
-    typedef struct omp_lock_t {
-        void * _lk;
-    } omp_lock_t;
-
-    extern void   __KAI_KMPC_CONVENTION  omp_init_lock    (omp_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_lock     (omp_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_unset_lock   (omp_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_destroy_lock (omp_lock_t *);
-    extern int    __KAI_KMPC_CONVENTION  omp_test_lock    (omp_lock_t *);
-
-    /* nested lock API functions */
-    typedef struct omp_nest_lock_t {
-        void * _lk;
-    } omp_nest_lock_t;
-
-    extern void   __KAI_KMPC_CONVENTION  omp_init_nest_lock    (omp_nest_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_set_nest_lock     (omp_nest_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_unset_nest_lock   (omp_nest_lock_t *);
-    extern void   __KAI_KMPC_CONVENTION  omp_destroy_nest_lock (omp_nest_lock_t *);
-    extern int    __KAI_KMPC_CONVENTION  omp_test_nest_lock    (omp_nest_lock_t *);
-
-    /* OpenMP 5.0  Synchronization hints*/
-    typedef enum omp_sync_hint_t {
-        omp_sync_hint_none           = 0,
-        omp_lock_hint_none           = omp_sync_hint_none,
-        omp_sync_hint_uncontended    = 1,
-        omp_lock_hint_uncontended    = omp_sync_hint_uncontended,
-        omp_sync_hint_contended      = (1<<1),
-        omp_lock_hint_contended      = omp_sync_hint_contended,
-        omp_sync_hint_nonspeculative = (1<<2),
-        omp_lock_hint_nonspeculative = omp_sync_hint_nonspeculative,
-        omp_sync_hint_speculative    = (1<<3),
-        omp_lock_hint_speculative    = omp_sync_hint_speculative,
-        kmp_lock_hint_hle            = (1<<16),
-        kmp_lock_hint_rtm            = (1<<17),
-        kmp_lock_hint_adaptive       = (1<<18)
-    } omp_sync_hint_t;
-
-    /* lock hint type for dynamic user lock */
-    typedef omp_sync_hint_t omp_lock_hint_t;
-
-    /* hinted lock initializers */
-    extern void __KAI_KMPC_CONVENTION omp_init_lock_with_hint(omp_lock_t *, omp_lock_hint_t);
-    extern void __KAI_KMPC_CONVENTION omp_init_nest_lock_with_hint(omp_nest_lock_t *, omp_lock_hint_t);
-
-    /* time API functions */
-    extern double __KAI_KMPC_CONVENTION  omp_get_wtime (void);
-    extern double __KAI_KMPC_CONVENTION  omp_get_wtick (void);
-
-    /* OpenMP 4.0 */
-    extern int  __KAI_KMPC_CONVENTION  omp_get_default_device (void);
-    extern void __KAI_KMPC_CONVENTION  omp_set_default_device (int);
-    extern int  __KAI_KMPC_CONVENTION  omp_is_initial_device (void);
-    extern int  __KAI_KMPC_CONVENTION  omp_get_num_devices (void);
-    extern int  __KAI_KMPC_CONVENTION  omp_get_num_teams (void);
-    extern int  __KAI_KMPC_CONVENTION  omp_get_team_num (void);
-    extern int  __KAI_KMPC_CONVENTION  omp_get_cancellation (void);
-
-#   include <stdlib.h>
-#   include <stdint.h>
-    /* OpenMP 4.5 */
-    extern int   __KAI_KMPC_CONVENTION  omp_get_initial_device (void);
-    extern void* __KAI_KMPC_CONVENTION  omp_target_alloc(size_t, int);
-    extern void  __KAI_KMPC_CONVENTION  omp_target_free(void *, int);
-    extern int   __KAI_KMPC_CONVENTION  omp_target_is_present(void *, int);
-    extern int   __KAI_KMPC_CONVENTION  omp_target_memcpy(void *, void *, size_t, size_t, size_t, int, int);
-    extern int   __KAI_KMPC_CONVENTION  omp_target_memcpy_rect(void *, void *, size_t, int, const size_t *,
-                                            const size_t *, const size_t *, const size_t *, const size_t *, int, int);
-    extern int   __KAI_KMPC_CONVENTION  omp_target_associate_ptr(void *, void *, size_t, size_t, int);
-    extern int   __KAI_KMPC_CONVENTION  omp_target_disassociate_ptr(void *, int);
-
-    /* OpenMP 5.0 */
-    extern int   __KAI_KMPC_CONVENTION  omp_get_device_num (void);
-    typedef void * omp_depend_t;
-
-    /* kmp API functions */
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_stacksize          (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_stacksize          (int);
-    extern size_t __KAI_KMPC_CONVENTION  kmp_get_stacksize_s        (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_stacksize_s        (size_t);
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_blocktime          (void);
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_library            (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_blocktime          (int);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_library            (int);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_library_serial     (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_library_turnaround (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_library_throughput (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_defaults           (char const *);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_disp_num_buffers   (int);
-
-    /* Intel affinity API */
-    typedef void * kmp_affinity_mask_t;
-
-    extern int    __KAI_KMPC_CONVENTION  kmp_set_affinity             (kmp_affinity_mask_t *);
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_affinity             (kmp_affinity_mask_t *);
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_affinity_max_proc    (void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_create_affinity_mask     (kmp_affinity_mask_t *);
-    extern void   __KAI_KMPC_CONVENTION  kmp_destroy_affinity_mask    (kmp_affinity_mask_t *);
-    extern int    __KAI_KMPC_CONVENTION  kmp_set_affinity_mask_proc   (int, kmp_affinity_mask_t *);
-    extern int    __KAI_KMPC_CONVENTION  kmp_unset_affinity_mask_proc (int, kmp_affinity_mask_t *);
-    extern int    __KAI_KMPC_CONVENTION  kmp_get_affinity_mask_proc   (int, kmp_affinity_mask_t *);
-
-    /* OpenMP 4.0 affinity API */
-    typedef enum omp_proc_bind_t {
-        omp_proc_bind_false = 0,
-        omp_proc_bind_true = 1,
-        omp_proc_bind_master = 2,
-        omp_proc_bind_close = 3,
-        omp_proc_bind_spread = 4
-    } omp_proc_bind_t;
-
-    extern omp_proc_bind_t __KAI_KMPC_CONVENTION omp_get_proc_bind (void);
-
-    /* OpenMP 4.5 affinity API */
-    extern int  __KAI_KMPC_CONVENTION omp_get_num_places (void);
-    extern int  __KAI_KMPC_CONVENTION omp_get_place_num_procs (int);
-    extern void __KAI_KMPC_CONVENTION omp_get_place_proc_ids (int, int *);
-    extern int  __KAI_KMPC_CONVENTION omp_get_place_num (void);
-    extern int  __KAI_KMPC_CONVENTION omp_get_partition_num_places (void);
-    extern void __KAI_KMPC_CONVENTION omp_get_partition_place_nums (int *);
-
-    extern void * __KAI_KMPC_CONVENTION  kmp_malloc  (size_t);
-    extern void * __KAI_KMPC_CONVENTION  kmp_aligned_malloc  (size_t, size_t);
-    extern void * __KAI_KMPC_CONVENTION  kmp_calloc  (size_t, size_t);
-    extern void * __KAI_KMPC_CONVENTION  kmp_realloc (void *, size_t);
-    extern void   __KAI_KMPC_CONVENTION  kmp_free    (void *);
-
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_warnings_on(void);
-    extern void   __KAI_KMPC_CONVENTION  kmp_set_warnings_off(void);
-
-    /* OpenMP 5.0 Tool Control */
-    typedef enum omp_control_tool_result_t {
-        omp_control_tool_notool = -2,
-        omp_control_tool_nocallback = -1,
-        omp_control_tool_success = 0,
-        omp_control_tool_ignored = 1
-    } omp_control_tool_result_t;
-
-    typedef enum omp_control_tool_t {
-        omp_control_tool_start = 1,
-        omp_control_tool_pause = 2,
-        omp_control_tool_flush = 3,
-        omp_control_tool_end = 4
-    } omp_control_tool_t;
-
-    extern int __KAI_KMPC_CONVENTION omp_control_tool(int, int, void*);
-
-    /* OpenMP 5.0 Memory Management */
-    typedef uintptr_t omp_uintptr_t;
-
-    typedef enum {
-        OMP_ATK_THREADMODEL = 1,
-        OMP_ATK_ALIGNMENT = 2,
-        OMP_ATK_ACCESS = 3,
-        OMP_ATK_POOL_SIZE = 4,
-        OMP_ATK_FALLBACK = 5,
-        OMP_ATK_FB_DATA = 6,
-        OMP_ATK_PINNED = 7,
-        OMP_ATK_PARTITION = 8
-    } omp_alloctrait_key_t;
-
-    typedef enum {
-        OMP_ATV_FALSE = 0,
-        OMP_ATV_TRUE = 1,
-        OMP_ATV_DEFAULT = 2,
-        OMP_ATV_CONTENDED = 3,
-        OMP_ATV_UNCONTENDED = 4,
-        OMP_ATV_SEQUENTIAL = 5,
-        OMP_ATV_PRIVATE = 6,
-        OMP_ATV_ALL = 7,
-        OMP_ATV_THREAD = 8,
-        OMP_ATV_PTEAM = 9,
-        OMP_ATV_CGROUP = 10,
-        OMP_ATV_DEFAULT_MEM_FB = 11,
-        OMP_ATV_NULL_FB = 12,
-        OMP_ATV_ABORT_FB = 13,
-        OMP_ATV_ALLOCATOR_FB = 14,
-        OMP_ATV_ENVIRONMENT = 15,
-        OMP_ATV_NEAREST = 16,
-        OMP_ATV_BLOCKED = 17,
-        OMP_ATV_INTERLEAVED = 18
-    } omp_alloctrait_value_t;
-
-    typedef struct {
-        omp_alloctrait_key_t key;
-        omp_uintptr_t value;
-    } omp_alloctrait_t;
-
-#   if defined(_WIN32)
-    // On Windows cl and icl do not support 64-bit enum, let's use integer then.
-    typedef omp_uintptr_t omp_allocator_handle_t;
-    extern __KMP_IMP omp_allocator_handle_t const omp_null_allocator;
-    extern __KMP_IMP omp_allocator_handle_t const omp_default_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_large_cap_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_const_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_high_bw_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_low_lat_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_cgroup_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_pteam_mem_alloc;
-    extern __KMP_IMP omp_allocator_handle_t const omp_thread_mem_alloc;
-    typedef omp_uintptr_t omp_memspace_handle_t;
-    extern __KMP_IMP omp_memspace_handle_t const omp_default_mem_space;
-    extern __KMP_IMP omp_memspace_handle_t const omp_large_cap_mem_space;
-    extern __KMP_IMP omp_memspace_handle_t const omp_const_mem_space;
-    extern __KMP_IMP omp_memspace_handle_t const omp_high_bw_mem_space;
-    extern __KMP_IMP omp_memspace_handle_t const omp_low_lat_mem_space;
-#   else
-#       if __cplusplus >= 201103
-    typedef enum omp_allocator_handle_t : omp_uintptr_t
-#       else
-    typedef enum omp_allocator_handle_t
-#       endif
-    {
-      omp_null_allocator = 0,
-      omp_default_mem_alloc = 1,
-      omp_large_cap_mem_alloc = 2,
-      omp_const_mem_alloc = 3,
-      omp_high_bw_mem_alloc = 4,
-      omp_low_lat_mem_alloc = 5,
-      omp_cgroup_mem_alloc = 6,
-      omp_pteam_mem_alloc = 7,
-      omp_thread_mem_alloc = 8,
-      KMP_ALLOCATOR_MAX_HANDLE = UINTPTR_MAX
-    } omp_allocator_handle_t;
-#       if __cplusplus >= 201103
-    typedef enum omp_memspace_handle_t : omp_uintptr_t
-#       else
-    typedef enum omp_memspace_handle_t
-#       endif
-    {
-      omp_default_mem_space = 0,
-      omp_large_cap_mem_space = 1,
-      omp_const_mem_space = 2,
-      omp_high_bw_mem_space = 3,
-      omp_low_lat_mem_space = 4,
-      KMP_MEMSPACE_MAX_HANDLE = UINTPTR_MAX
-    } omp_memspace_handle_t;
-#   endif
-    extern omp_allocator_handle_t __KAI_KMPC_CONVENTION omp_init_allocator(omp_memspace_handle_t m,
-                                                       int ntraits, omp_alloctrait_t traits[]);
-    extern void __KAI_KMPC_CONVENTION omp_destroy_allocator(omp_allocator_handle_t allocator);
-
-    extern void __KAI_KMPC_CONVENTION omp_set_default_allocator(omp_allocator_handle_t a);
-    extern omp_allocator_handle_t __KAI_KMPC_CONVENTION omp_get_default_allocator(void);
-#   ifdef __cplusplus
-    extern void *__KAI_KMPC_CONVENTION omp_alloc(size_t size, omp_allocator_handle_t a = omp_null_allocator);
-    extern void __KAI_KMPC_CONVENTION omp_free(void * ptr, omp_allocator_handle_t a = omp_null_allocator);
-#   else
-    extern void *__KAI_KMPC_CONVENTION omp_alloc(size_t size, omp_allocator_handle_t a);
-    extern void __KAI_KMPC_CONVENTION omp_free(void *ptr, omp_allocator_handle_t a);
-#   endif
-
-    /* OpenMP 5.0 Affinity Format */
-    extern void __KAI_KMPC_CONVENTION omp_set_affinity_format(char const *);
-    extern size_t __KAI_KMPC_CONVENTION omp_get_affinity_format(char *, size_t);
-    extern void __KAI_KMPC_CONVENTION omp_display_affinity(char const *);
-    extern size_t __KAI_KMPC_CONVENTION omp_capture_affinity(char *, size_t, char const *);
-
-    /* OpenMP 5.0 events */
-#   if defined(_WIN32)
-    // On Windows cl and icl do not support 64-bit enum, let's use integer then.
-    typedef omp_uintptr_t omp_event_handle_t;
-#   else
-    typedef enum omp_event_handle_t { KMP_EVENT_MAX_HANDLE = UINTPTR_MAX } omp_event_handle_t;
-#   endif
-    extern void __KAI_KMPC_CONVENTION omp_fulfill_event ( omp_event_handle_t event );
-
-    /* OpenMP 5.0 Pause Resources */
-    typedef enum omp_pause_resource_t {
-      omp_pause_resume = 0,
-      omp_pause_soft = 1,
-      omp_pause_hard = 2
-    } omp_pause_resource_t;
-    extern int __KAI_KMPC_CONVENTION omp_pause_resource(omp_pause_resource_t, int);
-    extern int __KAI_KMPC_CONVENTION omp_pause_resource_all(omp_pause_resource_t);
-
-    extern int __KAI_KMPC_CONVENTION omp_get_supported_active_levels(void);
-
-#   undef __KAI_KMPC_CONVENTION
-#   undef __KMP_IMP
-
-    /* Warning:
-       The following typedefs are not standard, deprecated and will be removed in a future release.
-    */
-    typedef int     omp_int_t;
-    typedef double  omp_wtime_t;
-
-#   ifdef __cplusplus
-    }
-#   endif
-
-#endif /* __OMP_H */
diff --git a/linux-x86/lib64/clang/9.0.7/include/opencl-c-base.h b/linux-x86/lib64/clang/9.0.7/include/opencl-c-base.h
deleted file mode 100644
index a82954d..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/opencl-c-base.h
+++ /dev/null
@@ -1,578 +0,0 @@
-//===----- opencl-c-base.h - OpenCL C language base definitions -----------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef _OPENCL_BASE_H_
-#define _OPENCL_BASE_H_
-
-// built-in scalar data types:
-
-/**
- * An unsigned 8-bit integer.
- */
-typedef unsigned char uchar;
-
-/**
- * An unsigned 16-bit integer.
- */
-typedef unsigned short ushort;
-
-/**
- * An unsigned 32-bit integer.
- */
-typedef unsigned int uint;
-
-/**
- * An unsigned 64-bit integer.
- */
-typedef unsigned long ulong;
-
-/**
- * The unsigned integer type of the result of the sizeof operator. This
- * is a 32-bit unsigned integer if CL_DEVICE_ADDRESS_BITS
- * defined in table 4.3 is 32-bits and is a 64-bit unsigned integer if
- * CL_DEVICE_ADDRESS_BITS is 64-bits.
- */
-typedef __SIZE_TYPE__ size_t;
-
-/**
- * A signed integer type that is the result of subtracting two pointers.
- * This is a 32-bit signed integer if CL_DEVICE_ADDRESS_BITS
- * defined in table 4.3 is 32-bits and is a 64-bit signed integer if
- * CL_DEVICE_ADDRESS_BITS is 64-bits.
- */
-typedef __PTRDIFF_TYPE__ ptrdiff_t;
-
-/**
- * A signed integer type with the property that any valid pointer to
- * void can be converted to this type, then converted back to pointer
- * to void, and the result will compare equal to the original pointer.
- */
-typedef __INTPTR_TYPE__ intptr_t;
-
-/**
- * An unsigned integer type with the property that any valid pointer to
- * void can be converted to this type, then converted back to pointer
- * to void, and the result will compare equal to the original pointer.
- */
-typedef __UINTPTR_TYPE__ uintptr_t;
-
-// built-in vector data types:
-typedef char char2 __attribute__((ext_vector_type(2)));
-typedef char char3 __attribute__((ext_vector_type(3)));
-typedef char char4 __attribute__((ext_vector_type(4)));
-typedef char char8 __attribute__((ext_vector_type(8)));
-typedef char char16 __attribute__((ext_vector_type(16)));
-typedef uchar uchar2 __attribute__((ext_vector_type(2)));
-typedef uchar uchar3 __attribute__((ext_vector_type(3)));
-typedef uchar uchar4 __attribute__((ext_vector_type(4)));
-typedef uchar uchar8 __attribute__((ext_vector_type(8)));
-typedef uchar uchar16 __attribute__((ext_vector_type(16)));
-typedef short short2 __attribute__((ext_vector_type(2)));
-typedef short short3 __attribute__((ext_vector_type(3)));
-typedef short short4 __attribute__((ext_vector_type(4)));
-typedef short short8 __attribute__((ext_vector_type(8)));
-typedef short short16 __attribute__((ext_vector_type(16)));
-typedef ushort ushort2 __attribute__((ext_vector_type(2)));
-typedef ushort ushort3 __attribute__((ext_vector_type(3)));
-typedef ushort ushort4 __attribute__((ext_vector_type(4)));
-typedef ushort ushort8 __attribute__((ext_vector_type(8)));
-typedef ushort ushort16 __attribute__((ext_vector_type(16)));
-typedef int int2 __attribute__((ext_vector_type(2)));
-typedef int int3 __attribute__((ext_vector_type(3)));
-typedef int int4 __attribute__((ext_vector_type(4)));
-typedef int int8 __attribute__((ext_vector_type(8)));
-typedef int int16 __attribute__((ext_vector_type(16)));
-typedef uint uint2 __attribute__((ext_vector_type(2)));
-typedef uint uint3 __attribute__((ext_vector_type(3)));
-typedef uint uint4 __attribute__((ext_vector_type(4)));
-typedef uint uint8 __attribute__((ext_vector_type(8)));
-typedef uint uint16 __attribute__((ext_vector_type(16)));
-typedef long long2 __attribute__((ext_vector_type(2)));
-typedef long long3 __attribute__((ext_vector_type(3)));
-typedef long long4 __attribute__((ext_vector_type(4)));
-typedef long long8 __attribute__((ext_vector_type(8)));
-typedef long long16 __attribute__((ext_vector_type(16)));
-typedef ulong ulong2 __attribute__((ext_vector_type(2)));
-typedef ulong ulong3 __attribute__((ext_vector_type(3)));
-typedef ulong ulong4 __attribute__((ext_vector_type(4)));
-typedef ulong ulong8 __attribute__((ext_vector_type(8)));
-typedef ulong ulong16 __attribute__((ext_vector_type(16)));
-typedef float float2 __attribute__((ext_vector_type(2)));
-typedef float float3 __attribute__((ext_vector_type(3)));
-typedef float float4 __attribute__((ext_vector_type(4)));
-typedef float float8 __attribute__((ext_vector_type(8)));
-typedef float float16 __attribute__((ext_vector_type(16)));
-#ifdef cl_khr_fp16
-#pragma OPENCL EXTENSION cl_khr_fp16 : enable
-typedef half half2 __attribute__((ext_vector_type(2)));
-typedef half half3 __attribute__((ext_vector_type(3)));
-typedef half half4 __attribute__((ext_vector_type(4)));
-typedef half half8 __attribute__((ext_vector_type(8)));
-typedef half half16 __attribute__((ext_vector_type(16)));
-#endif
-#ifdef cl_khr_fp64
-#if __OPENCL_C_VERSION__ < CL_VERSION_1_2
-#pragma OPENCL EXTENSION cl_khr_fp64 : enable
-#endif
-typedef double double2 __attribute__((ext_vector_type(2)));
-typedef double double3 __attribute__((ext_vector_type(3)));
-typedef double double4 __attribute__((ext_vector_type(4)));
-typedef double double8 __attribute__((ext_vector_type(8)));
-typedef double double16 __attribute__((ext_vector_type(16)));
-#endif
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#define NULL ((void*)0)
-#endif
-
-/**
- * Value of maximum non-infinite single-precision floating-point
- * number.
- */
-#define MAXFLOAT 0x1.fffffep127f
-
-/**
- * A positive float constant expression. HUGE_VALF evaluates
- * to +infinity. Used as an error value returned by the built-in
- * math functions.
- */
-#define HUGE_VALF (__builtin_huge_valf())
-
-/**
- * A positive double constant expression. HUGE_VAL evaluates
- * to +infinity. Used as an error value returned by the built-in
- * math functions.
- */
-#define HUGE_VAL (__builtin_huge_val())
-
-/**
- * A constant expression of type float representing positive or
- * unsigned infinity.
- */
-#define INFINITY (__builtin_inff())
-
-/**
- * A constant expression of type float representing a quiet NaN.
- */
-#define NAN as_float(INT_MAX)
-
-#define FP_ILOGB0    INT_MIN
-#define FP_ILOGBNAN  INT_MAX
-
-#define FLT_DIG 6
-#define FLT_MANT_DIG 24
-#define FLT_MAX_10_EXP +38
-#define FLT_MAX_EXP +128
-#define FLT_MIN_10_EXP -37
-#define FLT_MIN_EXP -125
-#define FLT_RADIX 2
-#define FLT_MAX 0x1.fffffep127f
-#define FLT_MIN 0x1.0p-126f
-#define FLT_EPSILON 0x1.0p-23f
-
-#define M_E_F         2.71828182845904523536028747135266250f
-#define M_LOG2E_F     1.44269504088896340735992468100189214f
-#define M_LOG10E_F    0.434294481903251827651128918916605082f
-#define M_LN2_F       0.693147180559945309417232121458176568f
-#define M_LN10_F      2.30258509299404568401799145468436421f
-#define M_PI_F        3.14159265358979323846264338327950288f
-#define M_PI_2_F      1.57079632679489661923132169163975144f
-#define M_PI_4_F      0.785398163397448309615660845819875721f
-#define M_1_PI_F      0.318309886183790671537767526745028724f
-#define M_2_PI_F      0.636619772367581343075535053490057448f
-#define M_2_SQRTPI_F  1.12837916709551257389615890312154517f
-#define M_SQRT2_F     1.41421356237309504880168872420969808f
-#define M_SQRT1_2_F   0.707106781186547524400844362104849039f
-
-#define DBL_DIG 15
-#define DBL_MANT_DIG 53
-#define DBL_MAX_10_EXP +308
-#define DBL_MAX_EXP +1024
-#define DBL_MIN_10_EXP -307
-#define DBL_MIN_EXP -1021
-#define DBL_RADIX 2
-#define DBL_MAX 0x1.fffffffffffffp1023
-#define DBL_MIN 0x1.0p-1022
-#define DBL_EPSILON 0x1.0p-52
-
-#define M_E           0x1.5bf0a8b145769p+1
-#define M_LOG2E       0x1.71547652b82fep+0
-#define M_LOG10E      0x1.bcb7b1526e50ep-2
-#define M_LN2         0x1.62e42fefa39efp-1
-#define M_LN10        0x1.26bb1bbb55516p+1
-#define M_PI          0x1.921fb54442d18p+1
-#define M_PI_2        0x1.921fb54442d18p+0
-#define M_PI_4        0x1.921fb54442d18p-1
-#define M_1_PI        0x1.45f306dc9c883p-2
-#define M_2_PI        0x1.45f306dc9c883p-1
-#define M_2_SQRTPI    0x1.20dd750429b6dp+0
-#define M_SQRT2       0x1.6a09e667f3bcdp+0
-#define M_SQRT1_2     0x1.6a09e667f3bcdp-1
-
-#ifdef cl_khr_fp16
-
-#define HALF_DIG 3
-#define HALF_MANT_DIG 11
-#define HALF_MAX_10_EXP +4
-#define HALF_MAX_EXP +16
-#define HALF_MIN_10_EXP -4
-#define HALF_MIN_EXP -13
-#define HALF_RADIX 2
-#define HALF_MAX ((0x1.ffcp15h))
-#define HALF_MIN ((0x1.0p-14h))
-#define HALF_EPSILON ((0x1.0p-10h))
-
-#define M_E_H         2.71828182845904523536028747135266250h
-#define M_LOG2E_H     1.44269504088896340735992468100189214h
-#define M_LOG10E_H    0.434294481903251827651128918916605082h
-#define M_LN2_H       0.693147180559945309417232121458176568h
-#define M_LN10_H      2.30258509299404568401799145468436421h
-#define M_PI_H        3.14159265358979323846264338327950288h
-#define M_PI_2_H      1.57079632679489661923132169163975144h
-#define M_PI_4_H      0.785398163397448309615660845819875721h
-#define M_1_PI_H      0.318309886183790671537767526745028724h
-#define M_2_PI_H      0.636619772367581343075535053490057448h
-#define M_2_SQRTPI_H  1.12837916709551257389615890312154517h
-#define M_SQRT2_H     1.41421356237309504880168872420969808h
-#define M_SQRT1_2_H   0.707106781186547524400844362104849039h
-
-#endif //cl_khr_fp16
-
-#define CHAR_BIT  8
-#define SCHAR_MAX 127
-#define SCHAR_MIN (-128)
-#define UCHAR_MAX 255
-#define CHAR_MAX  SCHAR_MAX
-#define CHAR_MIN  SCHAR_MIN
-#define USHRT_MAX 65535
-#define SHRT_MAX  32767
-#define SHRT_MIN  (-32768)
-#define UINT_MAX  0xffffffff
-#define INT_MAX   2147483647
-#define INT_MIN   (-2147483647-1)
-#define ULONG_MAX 0xffffffffffffffffUL
-#define LONG_MAX  0x7fffffffffffffffL
-#define LONG_MIN  (-0x7fffffffffffffffL-1)
-
-// OpenCL v1.1 s6.11.8, v1.2 s6.12.8, v2.0 s6.13.8 - Synchronization Functions
-
-// Flag type and values for barrier, mem_fence, read_mem_fence, write_mem_fence
-typedef uint cl_mem_fence_flags;
-
-/**
- * Queue a memory fence to ensure correct
- * ordering of memory operations to local memory
- */
-#define CLK_LOCAL_MEM_FENCE    0x01
-
-/**
- * Queue a memory fence to ensure correct
- * ordering of memory operations to global memory
- */
-#define CLK_GLOBAL_MEM_FENCE   0x02
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-typedef enum memory_scope {
-  memory_scope_work_item = __OPENCL_MEMORY_SCOPE_WORK_ITEM,
-  memory_scope_work_group = __OPENCL_MEMORY_SCOPE_WORK_GROUP,
-  memory_scope_device = __OPENCL_MEMORY_SCOPE_DEVICE,
-  memory_scope_all_svm_devices = __OPENCL_MEMORY_SCOPE_ALL_SVM_DEVICES,
-#if defined(cl_intel_subgroups) || defined(cl_khr_subgroups)
-  memory_scope_sub_group = __OPENCL_MEMORY_SCOPE_SUB_GROUP
-#endif
-} memory_scope;
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-/**
- * Queue a memory fence to ensure correct ordering of memory
- * operations between work-items of a work-group to
- * image memory.
- */
-#define CLK_IMAGE_MEM_FENCE  0x04
-
-#ifndef ATOMIC_VAR_INIT
-#define ATOMIC_VAR_INIT(x) (x)
-#endif //ATOMIC_VAR_INIT
-#define ATOMIC_FLAG_INIT 0
-
-// enum values aligned with what clang uses in EmitAtomicExpr()
-typedef enum memory_order
-{
-  memory_order_relaxed = __ATOMIC_RELAXED,
-  memory_order_acquire = __ATOMIC_ACQUIRE,
-  memory_order_release = __ATOMIC_RELEASE,
-  memory_order_acq_rel = __ATOMIC_ACQ_REL,
-  memory_order_seq_cst = __ATOMIC_SEQ_CST
-} memory_order;
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.3, v1.2 s6.12.14, v2.0 s6.13.14 - Image Read and Write Functions
-
-// These values need to match the runtime equivalent
-//
-// Addressing Mode.
-//
-#define CLK_ADDRESS_NONE                0
-#define CLK_ADDRESS_CLAMP_TO_EDGE       2
-#define CLK_ADDRESS_CLAMP               4
-#define CLK_ADDRESS_REPEAT              6
-#define CLK_ADDRESS_MIRRORED_REPEAT     8
-
-//
-// Coordination Normalization
-//
-#define CLK_NORMALIZED_COORDS_FALSE     0
-#define CLK_NORMALIZED_COORDS_TRUE      1
-
-//
-// Filtering Mode.
-//
-#define CLK_FILTER_NEAREST              0x10
-#define CLK_FILTER_LINEAR               0x20
-
-#ifdef cl_khr_gl_msaa_sharing
-#pragma OPENCL EXTENSION cl_khr_gl_msaa_sharing : enable
-#endif //cl_khr_gl_msaa_sharing
-
-//
-// Channel Datatype.
-//
-#define CLK_SNORM_INT8        0x10D0
-#define CLK_SNORM_INT16       0x10D1
-#define CLK_UNORM_INT8        0x10D2
-#define CLK_UNORM_INT16       0x10D3
-#define CLK_UNORM_SHORT_565   0x10D4
-#define CLK_UNORM_SHORT_555   0x10D5
-#define CLK_UNORM_INT_101010  0x10D6
-#define CLK_SIGNED_INT8       0x10D7
-#define CLK_SIGNED_INT16      0x10D8
-#define CLK_SIGNED_INT32      0x10D9
-#define CLK_UNSIGNED_INT8     0x10DA
-#define CLK_UNSIGNED_INT16    0x10DB
-#define CLK_UNSIGNED_INT32    0x10DC
-#define CLK_HALF_FLOAT        0x10DD
-#define CLK_FLOAT             0x10DE
-#define CLK_UNORM_INT24       0x10DF
-
-// Channel order, numbering must be aligned with cl_channel_order in cl.h
-//
-#define CLK_R         0x10B0
-#define CLK_A         0x10B1
-#define CLK_RG        0x10B2
-#define CLK_RA        0x10B3
-#define CLK_RGB       0x10B4
-#define CLK_RGBA      0x10B5
-#define CLK_BGRA      0x10B6
-#define CLK_ARGB      0x10B7
-#define CLK_INTENSITY 0x10B8
-#define CLK_LUMINANCE 0x10B9
-#define CLK_Rx                0x10BA
-#define CLK_RGx               0x10BB
-#define CLK_RGBx              0x10BC
-#define CLK_DEPTH             0x10BD
-#define CLK_DEPTH_STENCIL     0x10BE
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#define CLK_sRGB              0x10BF
-#define CLK_sRGBx             0x10C0
-#define CLK_sRGBA             0x10C1
-#define CLK_sBGRA             0x10C2
-#define CLK_ABGR              0x10C3
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#define CLK_NULL_RESERVE_ID (__builtin_astype(((void*)(__SIZE_MAX__)), reserve_id_t))
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-
-// OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#define CL_COMPLETE                                 0x0
-#define CL_RUNNING                                  0x1
-#define CL_SUBMITTED                                0x2
-#define CL_QUEUED                                   0x3
-
-#define CLK_SUCCESS                                 0
-#define CLK_ENQUEUE_FAILURE                         -101
-#define CLK_INVALID_QUEUE                           -102
-#define CLK_INVALID_NDRANGE                         -160
-#define CLK_INVALID_EVENT_WAIT_LIST                 -57
-#define CLK_DEVICE_QUEUE_FULL                       -161
-#define CLK_INVALID_ARG_SIZE                        -51
-#define CLK_EVENT_ALLOCATION_FAILURE                -100
-#define CLK_OUT_OF_RESOURCES                        -5
-
-#define CLK_NULL_QUEUE                              0
-#define CLK_NULL_EVENT (__builtin_astype(((void*)(__SIZE_MAX__)), clk_event_t))
-
-// execution model related definitions
-#define CLK_ENQUEUE_FLAGS_NO_WAIT                   0x0
-#define CLK_ENQUEUE_FLAGS_WAIT_KERNEL               0x1
-#define CLK_ENQUEUE_FLAGS_WAIT_WORK_GROUP           0x2
-
-typedef int kernel_enqueue_flags_t;
-typedef int clk_profiling_info;
-
-// Profiling info name (see capture_event_profiling_info)
-#define CLK_PROFILING_COMMAND_EXEC_TIME 0x1
-
-#define MAX_WORK_DIM 3
-
-typedef struct {
-  unsigned int workDimension;
-  size_t globalWorkOffset[MAX_WORK_DIM];
-  size_t globalWorkSize[MAX_WORK_DIM];
-  size_t localWorkSize[MAX_WORK_DIM];
-} ndrange_t;
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#ifdef cl_intel_device_side_avc_motion_estimation
-#pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : begin
-
-#define CLK_AVC_ME_MAJOR_16x16_INTEL 0x0
-#define CLK_AVC_ME_MAJOR_16x8_INTEL 0x1
-#define CLK_AVC_ME_MAJOR_8x16_INTEL 0x2
-#define CLK_AVC_ME_MAJOR_8x8_INTEL 0x3
-
-#define CLK_AVC_ME_MINOR_8x8_INTEL 0x0
-#define CLK_AVC_ME_MINOR_8x4_INTEL 0x1
-#define CLK_AVC_ME_MINOR_4x8_INTEL 0x2
-#define CLK_AVC_ME_MINOR_4x4_INTEL 0x3
-
-#define CLK_AVC_ME_MAJOR_FORWARD_INTEL 0x0
-#define CLK_AVC_ME_MAJOR_BACKWARD_INTEL 0x1
-#define CLK_AVC_ME_MAJOR_BIDIRECTIONAL_INTEL 0x2
-
-#define CLK_AVC_ME_PARTITION_MASK_ALL_INTEL 0x0
-#define CLK_AVC_ME_PARTITION_MASK_16x16_INTEL 0x7E
-#define CLK_AVC_ME_PARTITION_MASK_16x8_INTEL 0x7D
-#define CLK_AVC_ME_PARTITION_MASK_8x16_INTEL 0x7B
-#define CLK_AVC_ME_PARTITION_MASK_8x8_INTEL 0x77
-#define CLK_AVC_ME_PARTITION_MASK_8x4_INTEL 0x6F
-#define CLK_AVC_ME_PARTITION_MASK_4x8_INTEL 0x5F
-#define CLK_AVC_ME_PARTITION_MASK_4x4_INTEL 0x3F
-
-#define CLK_AVC_ME_SLICE_TYPE_PRED_INTEL 0x0
-#define CLK_AVC_ME_SLICE_TYPE_BPRED_INTEL 0x1
-#define CLK_AVC_ME_SLICE_TYPE_INTRA_INTEL 0x2
-
-#define CLK_AVC_ME_SEARCH_WINDOW_EXHAUSTIVE_INTEL 0x0
-#define CLK_AVC_ME_SEARCH_WINDOW_SMALL_INTEL 0x1
-#define CLK_AVC_ME_SEARCH_WINDOW_TINY_INTEL 0x2
-#define CLK_AVC_ME_SEARCH_WINDOW_EXTRA_TINY_INTEL 0x3
-#define CLK_AVC_ME_SEARCH_WINDOW_DIAMOND_INTEL 0x4
-#define CLK_AVC_ME_SEARCH_WINDOW_LARGE_DIAMOND_INTEL 0x5
-#define CLK_AVC_ME_SEARCH_WINDOW_RESERVED0_INTEL 0x6
-#define CLK_AVC_ME_SEARCH_WINDOW_RESERVED1_INTEL 0x7
-#define CLK_AVC_ME_SEARCH_WINDOW_CUSTOM_INTEL 0x8
-
-#define CLK_AVC_ME_SAD_ADJUST_MODE_NONE_INTEL 0x0
-#define CLK_AVC_ME_SAD_ADJUST_MODE_HAAR_INTEL 0x2
-
-#define CLK_AVC_ME_SUBPIXEL_MODE_INTEGER_INTEL 0x0
-#define CLK_AVC_ME_SUBPIXEL_MODE_HPEL_INTEL 0x1
-#define CLK_AVC_ME_SUBPIXEL_MODE_QPEL_INTEL 0x3
-
-#define CLK_AVC_ME_COST_PRECISION_QPEL_INTEL 0x0
-#define CLK_AVC_ME_COST_PRECISION_HPEL_INTEL 0x1
-#define CLK_AVC_ME_COST_PRECISION_PEL_INTEL 0x2
-#define CLK_AVC_ME_COST_PRECISION_DPEL_INTEL 0x3
-
-#define CLK_AVC_ME_BIDIR_WEIGHT_QUARTER_INTEL 0x10
-#define CLK_AVC_ME_BIDIR_WEIGHT_THIRD_INTEL 0x15
-#define CLK_AVC_ME_BIDIR_WEIGHT_HALF_INTEL 0x20
-#define CLK_AVC_ME_BIDIR_WEIGHT_TWO_THIRD_INTEL 0x2B
-#define CLK_AVC_ME_BIDIR_WEIGHT_THREE_QUARTER_INTEL 0x30
-
-#define CLK_AVC_ME_BORDER_REACHED_LEFT_INTEL 0x0
-#define CLK_AVC_ME_BORDER_REACHED_RIGHT_INTEL 0x2
-#define CLK_AVC_ME_BORDER_REACHED_TOP_INTEL 0x4
-#define CLK_AVC_ME_BORDER_REACHED_BOTTOM_INTEL 0x8
-
-#define CLK_AVC_ME_INTRA_16x16_INTEL 0x0
-#define CLK_AVC_ME_INTRA_8x8_INTEL 0x1
-#define CLK_AVC_ME_INTRA_4x4_INTEL 0x2
-
-#define CLK_AVC_ME_SKIP_BLOCK_PARTITION_16x16_INTEL 0x0
-#define CLK_AVC_ME_SKIP_BLOCK_PARTITION_8x8_INTEL 0x4000
-
-#define CLK_AVC_ME_SKIP_BLOCK_16x16_FORWARD_ENABLE_INTEL (0x1 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_16x16_BACKWARD_ENABLE_INTEL (0x2 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_16x16_DUAL_ENABLE_INTEL (0x3 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_FORWARD_ENABLE_INTEL (0x55 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_BACKWARD_ENABLE_INTEL (0xAA << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_DUAL_ENABLE_INTEL (0xFF << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_0_FORWARD_ENABLE_INTEL (0x1 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_0_BACKWARD_ENABLE_INTEL (0x2 << 24)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_1_FORWARD_ENABLE_INTEL (0x1 << 26)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_1_BACKWARD_ENABLE_INTEL (0x2 << 26)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_2_FORWARD_ENABLE_INTEL (0x1 << 28)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_2_BACKWARD_ENABLE_INTEL (0x2 << 28)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_3_FORWARD_ENABLE_INTEL (0x1 << 30)
-#define CLK_AVC_ME_SKIP_BLOCK_8x8_3_BACKWARD_ENABLE_INTEL (0x2 << 30)
-
-#define CLK_AVC_ME_BLOCK_BASED_SKIP_4x4_INTEL 0x00
-#define CLK_AVC_ME_BLOCK_BASED_SKIP_8x8_INTEL 0x80
-
-#define CLK_AVC_ME_INTRA_LUMA_PARTITION_MASK_ALL_INTEL 0x0
-#define CLK_AVC_ME_INTRA_LUMA_PARTITION_MASK_16x16_INTEL 0x6
-#define CLK_AVC_ME_INTRA_LUMA_PARTITION_MASK_8x8_INTEL 0x5
-#define CLK_AVC_ME_INTRA_LUMA_PARTITION_MASK_4x4_INTEL 0x3
-
-#define CLK_AVC_ME_INTRA_NEIGHBOR_LEFT_MASK_ENABLE_INTEL 0x60
-#define CLK_AVC_ME_INTRA_NEIGHBOR_UPPER_MASK_ENABLE_INTEL 0x10
-#define CLK_AVC_ME_INTRA_NEIGHBOR_UPPER_RIGHT_MASK_ENABLE_INTEL 0x8
-#define CLK_AVC_ME_INTRA_NEIGHBOR_UPPER_LEFT_MASK_ENABLE_INTEL 0x4
-
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_VERTICAL_INTEL 0x0
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_HORIZONTAL_INTEL 0x1
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_DC_INTEL 0x2
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_DIAGONAL_DOWN_LEFT_INTEL 0x3
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_DIAGONAL_DOWN_RIGHT_INTEL 0x4
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_PLANE_INTEL 0x4
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_VERTICAL_RIGHT_INTEL 0x5
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_HORIZONTAL_DOWN_INTEL 0x6
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_VERTICAL_LEFT_INTEL 0x7
-#define CLK_AVC_ME_LUMA_PREDICTOR_MODE_HORIZONTAL_UP_INTEL 0x8
-#define CLK_AVC_ME_CHROMA_PREDICTOR_MODE_DC_INTEL 0x0
-#define CLK_AVC_ME_CHROMA_PREDICTOR_MODE_HORIZONTAL_INTEL 0x1
-#define CLK_AVC_ME_CHROMA_PREDICTOR_MODE_VERTICAL_INTEL 0x2
-#define CLK_AVC_ME_CHROMA_PREDICTOR_MODE_PLANE_INTEL 0x3
-
-#define CLK_AVC_ME_FRAME_FORWARD_INTEL 0x1
-#define CLK_AVC_ME_FRAME_BACKWARD_INTEL 0x2
-#define CLK_AVC_ME_FRAME_DUAL_INTEL 0x3
-
-#define CLK_AVC_ME_INTERLACED_SCAN_TOP_FIELD_INTEL 0x0
-#define CLK_AVC_ME_INTERLACED_SCAN_BOTTOM_FIELD_INTEL 0x1
-
-#define CLK_AVC_ME_INITIALIZE_INTEL 0x0
-
-#define CLK_AVC_IME_PAYLOAD_INITIALIZE_INTEL 0x0
-#define CLK_AVC_REF_PAYLOAD_INITIALIZE_INTEL 0x0
-#define CLK_AVC_SIC_PAYLOAD_INITIALIZE_INTEL 0x0
-
-#define CLK_AVC_IME_RESULT_INITIALIZE_INTEL 0x0
-#define CLK_AVC_REF_RESULT_INITIALIZE_INTEL 0x0
-#define CLK_AVC_SIC_RESULT_INITIALIZE_INTEL 0x0
-
-#define CLK_AVC_IME_RESULT_SINGLE_REFERENCE_STREAMOUT_INITIALIZE_INTEL 0x0
-#define CLK_AVC_IME_RESULT_SINGLE_REFERENCE_STREAMIN_INITIALIZE_INTEL 0x0
-#define CLK_AVC_IME_RESULT_DUAL_REFERENCE_STREAMOUT_INITIALIZE_INTEL 0x0
-#define CLK_AVC_IME_RESULT_DUAL_REFERENCE_STREAMIN_INITIALIZE_INTEL 0x0
-
-#pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : end
-#endif // cl_intel_device_side_avc_motion_estimation
-
-#endif //_OPENCL_BASE_H_
diff --git a/linux-x86/lib64/clang/9.0.7/include/opencl-c.h b/linux-x86/lib64/clang/9.0.7/include/opencl-c.h
deleted file mode 100644
index 4207c53..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/opencl-c.h
+++ /dev/null
@@ -1,16502 +0,0 @@
-//===--- opencl-c.h - OpenCL C language builtin function header -----------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef _OPENCL_H_
-#define _OPENCL_H_
-
-#include "opencl-c-base.h"
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifndef cl_khr_depth_images
-#define cl_khr_depth_images
-#endif //cl_khr_depth_images
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
-#ifdef cl_khr_3d_image_writes
-#pragma OPENCL EXTENSION cl_khr_3d_image_writes : enable
-#endif //cl_khr_3d_image_writes
-#endif //__OPENCL_C_VERSION__ < CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-#pragma OPENCL EXTENSION cl_intel_planar_yuv : begin
-#pragma OPENCL EXTENSION cl_intel_planar_yuv : end
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-
-#define __ovld __attribute__((overloadable))
-#define __conv __attribute__((convergent))
-
-// Optimizations
-#define __purefn __attribute__((pure))
-#define __cnfn __attribute__((const))
-
-
-// OpenCL v1.1/1.2/2.0 s6.2.3 - Explicit conversions
-
-char __ovld __cnfn convert_char_rte(char);
-char __ovld __cnfn convert_char_sat_rte(char);
-char __ovld __cnfn convert_char_rtz(char);
-char __ovld __cnfn convert_char_sat_rtz(char);
-char __ovld __cnfn convert_char_rtp(char);
-char __ovld __cnfn convert_char_sat_rtp(char);
-char __ovld __cnfn convert_char_rtn(char);
-char __ovld __cnfn convert_char_sat_rtn(char);
-char __ovld __cnfn convert_char(char);
-char __ovld __cnfn convert_char_sat(char);
-char __ovld __cnfn convert_char_rte(uchar);
-char __ovld __cnfn convert_char_sat_rte(uchar);
-char __ovld __cnfn convert_char_rtz(uchar);
-char __ovld __cnfn convert_char_sat_rtz(uchar);
-char __ovld __cnfn convert_char_rtp(uchar);
-char __ovld __cnfn convert_char_sat_rtp(uchar);
-char __ovld __cnfn convert_char_rtn(uchar);
-char __ovld __cnfn convert_char_sat_rtn(uchar);
-char __ovld __cnfn convert_char(uchar);
-char __ovld __cnfn convert_char_sat(uchar);
-char __ovld __cnfn convert_char_rte(short);
-char __ovld __cnfn convert_char_sat_rte(short);
-char __ovld __cnfn convert_char_rtz(short);
-char __ovld __cnfn convert_char_sat_rtz(short);
-char __ovld __cnfn convert_char_rtp(short);
-char __ovld __cnfn convert_char_sat_rtp(short);
-char __ovld __cnfn convert_char_rtn(short);
-char __ovld __cnfn convert_char_sat_rtn(short);
-char __ovld __cnfn convert_char(short);
-char __ovld __cnfn convert_char_sat(short);
-char __ovld __cnfn convert_char_rte(ushort);
-char __ovld __cnfn convert_char_sat_rte(ushort);
-char __ovld __cnfn convert_char_rtz(ushort);
-char __ovld __cnfn convert_char_sat_rtz(ushort);
-char __ovld __cnfn convert_char_rtp(ushort);
-char __ovld __cnfn convert_char_sat_rtp(ushort);
-char __ovld __cnfn convert_char_rtn(ushort);
-char __ovld __cnfn convert_char_sat_rtn(ushort);
-char __ovld __cnfn convert_char(ushort);
-char __ovld __cnfn convert_char_sat(ushort);
-char __ovld __cnfn convert_char_rte(int);
-char __ovld __cnfn convert_char_sat_rte(int);
-char __ovld __cnfn convert_char_rtz(int);
-char __ovld __cnfn convert_char_sat_rtz(int);
-char __ovld __cnfn convert_char_rtp(int);
-char __ovld __cnfn convert_char_sat_rtp(int);
-char __ovld __cnfn convert_char_rtn(int);
-char __ovld __cnfn convert_char_sat_rtn(int);
-char __ovld __cnfn convert_char(int);
-char __ovld __cnfn convert_char_sat(int);
-char __ovld __cnfn convert_char_rte(uint);
-char __ovld __cnfn convert_char_sat_rte(uint);
-char __ovld __cnfn convert_char_rtz(uint);
-char __ovld __cnfn convert_char_sat_rtz(uint);
-char __ovld __cnfn convert_char_rtp(uint);
-char __ovld __cnfn convert_char_sat_rtp(uint);
-char __ovld __cnfn convert_char_rtn(uint);
-char __ovld __cnfn convert_char_sat_rtn(uint);
-char __ovld __cnfn convert_char(uint);
-char __ovld __cnfn convert_char_sat(uint);
-char __ovld __cnfn convert_char_rte(long);
-char __ovld __cnfn convert_char_sat_rte(long);
-char __ovld __cnfn convert_char_rtz(long);
-char __ovld __cnfn convert_char_sat_rtz(long);
-char __ovld __cnfn convert_char_rtp(long);
-char __ovld __cnfn convert_char_sat_rtp(long);
-char __ovld __cnfn convert_char_rtn(long);
-char __ovld __cnfn convert_char_sat_rtn(long);
-char __ovld __cnfn convert_char(long);
-char __ovld __cnfn convert_char_sat(long);
-char __ovld __cnfn convert_char_rte(ulong);
-char __ovld __cnfn convert_char_sat_rte(ulong);
-char __ovld __cnfn convert_char_rtz(ulong);
-char __ovld __cnfn convert_char_sat_rtz(ulong);
-char __ovld __cnfn convert_char_rtp(ulong);
-char __ovld __cnfn convert_char_sat_rtp(ulong);
-char __ovld __cnfn convert_char_rtn(ulong);
-char __ovld __cnfn convert_char_sat_rtn(ulong);
-char __ovld __cnfn convert_char(ulong);
-char __ovld __cnfn convert_char_sat(ulong);
-char __ovld __cnfn convert_char_rte(float);
-char __ovld __cnfn convert_char_sat_rte(float);
-char __ovld __cnfn convert_char_rtz(float);
-char __ovld __cnfn convert_char_sat_rtz(float);
-char __ovld __cnfn convert_char_rtp(float);
-char __ovld __cnfn convert_char_sat_rtp(float);
-char __ovld __cnfn convert_char_rtn(float);
-char __ovld __cnfn convert_char_sat_rtn(float);
-char __ovld __cnfn convert_char(float);
-char __ovld __cnfn convert_char_sat(float);
-uchar __ovld __cnfn convert_uchar_rte(char);
-uchar __ovld __cnfn convert_uchar_sat_rte(char);
-uchar __ovld __cnfn convert_uchar_rtz(char);
-uchar __ovld __cnfn convert_uchar_sat_rtz(char);
-uchar __ovld __cnfn convert_uchar_rtp(char);
-uchar __ovld __cnfn convert_uchar_sat_rtp(char);
-uchar __ovld __cnfn convert_uchar_rtn(char);
-uchar __ovld __cnfn convert_uchar_sat_rtn(char);
-uchar __ovld __cnfn convert_uchar(char);
-uchar __ovld __cnfn convert_uchar_sat(char);
-uchar __ovld __cnfn convert_uchar_rte(uchar);
-uchar __ovld __cnfn convert_uchar_sat_rte(uchar);
-uchar __ovld __cnfn convert_uchar_rtz(uchar);
-uchar __ovld __cnfn convert_uchar_sat_rtz(uchar);
-uchar __ovld __cnfn convert_uchar_rtp(uchar);
-uchar __ovld __cnfn convert_uchar_sat_rtp(uchar);
-uchar __ovld __cnfn convert_uchar_rtn(uchar);
-uchar __ovld __cnfn convert_uchar_sat_rtn(uchar);
-uchar __ovld __cnfn convert_uchar(uchar);
-uchar __ovld __cnfn convert_uchar_sat(uchar);
-uchar __ovld __cnfn convert_uchar_rte(short);
-uchar __ovld __cnfn convert_uchar_sat_rte(short);
-uchar __ovld __cnfn convert_uchar_rtz(short);
-uchar __ovld __cnfn convert_uchar_sat_rtz(short);
-uchar __ovld __cnfn convert_uchar_rtp(short);
-uchar __ovld __cnfn convert_uchar_sat_rtp(short);
-uchar __ovld __cnfn convert_uchar_rtn(short);
-uchar __ovld __cnfn convert_uchar_sat_rtn(short);
-uchar __ovld __cnfn convert_uchar(short);
-uchar __ovld __cnfn convert_uchar_sat(short);
-uchar __ovld __cnfn convert_uchar_rte(ushort);
-uchar __ovld __cnfn convert_uchar_sat_rte(ushort);
-uchar __ovld __cnfn convert_uchar_rtz(ushort);
-uchar __ovld __cnfn convert_uchar_sat_rtz(ushort);
-uchar __ovld __cnfn convert_uchar_rtp(ushort);
-uchar __ovld __cnfn convert_uchar_sat_rtp(ushort);
-uchar __ovld __cnfn convert_uchar_rtn(ushort);
-uchar __ovld __cnfn convert_uchar_sat_rtn(ushort);
-uchar __ovld __cnfn convert_uchar(ushort);
-uchar __ovld __cnfn convert_uchar_sat(ushort);
-uchar __ovld __cnfn convert_uchar_rte(int);
-uchar __ovld __cnfn convert_uchar_sat_rte(int);
-uchar __ovld __cnfn convert_uchar_rtz(int);
-uchar __ovld __cnfn convert_uchar_sat_rtz(int);
-uchar __ovld __cnfn convert_uchar_rtp(int);
-uchar __ovld __cnfn convert_uchar_sat_rtp(int);
-uchar __ovld __cnfn convert_uchar_rtn(int);
-uchar __ovld __cnfn convert_uchar_sat_rtn(int);
-uchar __ovld __cnfn convert_uchar(int);
-uchar __ovld __cnfn convert_uchar_sat(int);
-uchar __ovld __cnfn convert_uchar_rte(uint);
-uchar __ovld __cnfn convert_uchar_sat_rte(uint);
-uchar __ovld __cnfn convert_uchar_rtz(uint);
-uchar __ovld __cnfn convert_uchar_sat_rtz(uint);
-uchar __ovld __cnfn convert_uchar_rtp(uint);
-uchar __ovld __cnfn convert_uchar_sat_rtp(uint);
-uchar __ovld __cnfn convert_uchar_rtn(uint);
-uchar __ovld __cnfn convert_uchar_sat_rtn(uint);
-uchar __ovld __cnfn convert_uchar(uint);
-uchar __ovld __cnfn convert_uchar_sat(uint);
-uchar __ovld __cnfn convert_uchar_rte(long);
-uchar __ovld __cnfn convert_uchar_sat_rte(long);
-uchar __ovld __cnfn convert_uchar_rtz(long);
-uchar __ovld __cnfn convert_uchar_sat_rtz(long);
-uchar __ovld __cnfn convert_uchar_rtp(long);
-uchar __ovld __cnfn convert_uchar_sat_rtp(long);
-uchar __ovld __cnfn convert_uchar_rtn(long);
-uchar __ovld __cnfn convert_uchar_sat_rtn(long);
-uchar __ovld __cnfn convert_uchar(long);
-uchar __ovld __cnfn convert_uchar_sat(long);
-uchar __ovld __cnfn convert_uchar_rte(ulong);
-uchar __ovld __cnfn convert_uchar_sat_rte(ulong);
-uchar __ovld __cnfn convert_uchar_rtz(ulong);
-uchar __ovld __cnfn convert_uchar_sat_rtz(ulong);
-uchar __ovld __cnfn convert_uchar_rtp(ulong);
-uchar __ovld __cnfn convert_uchar_sat_rtp(ulong);
-uchar __ovld __cnfn convert_uchar_rtn(ulong);
-uchar __ovld __cnfn convert_uchar_sat_rtn(ulong);
-uchar __ovld __cnfn convert_uchar(ulong);
-uchar __ovld __cnfn convert_uchar_sat(ulong);
-uchar __ovld __cnfn convert_uchar_rte(float);
-uchar __ovld __cnfn convert_uchar_sat_rte(float);
-uchar __ovld __cnfn convert_uchar_rtz(float);
-uchar __ovld __cnfn convert_uchar_sat_rtz(float);
-uchar __ovld __cnfn convert_uchar_rtp(float);
-uchar __ovld __cnfn convert_uchar_sat_rtp(float);
-uchar __ovld __cnfn convert_uchar_rtn(float);
-uchar __ovld __cnfn convert_uchar_sat_rtn(float);
-uchar __ovld __cnfn convert_uchar(float);
-uchar __ovld __cnfn convert_uchar_sat(float);
-
-short __ovld __cnfn convert_short_rte(char);
-short __ovld __cnfn convert_short_sat_rte(char);
-short __ovld __cnfn convert_short_rtz(char);
-short __ovld __cnfn convert_short_sat_rtz(char);
-short __ovld __cnfn convert_short_rtp(char);
-short __ovld __cnfn convert_short_sat_rtp(char);
-short __ovld __cnfn convert_short_rtn(char);
-short __ovld __cnfn convert_short_sat_rtn(char);
-short __ovld __cnfn convert_short(char);
-short __ovld __cnfn convert_short_sat(char);
-short __ovld __cnfn convert_short_rte(uchar);
-short __ovld __cnfn convert_short_sat_rte(uchar);
-short __ovld __cnfn convert_short_rtz(uchar);
-short __ovld __cnfn convert_short_sat_rtz(uchar);
-short __ovld __cnfn convert_short_rtp(uchar);
-short __ovld __cnfn convert_short_sat_rtp(uchar);
-short __ovld __cnfn convert_short_rtn(uchar);
-short __ovld __cnfn convert_short_sat_rtn(uchar);
-short __ovld __cnfn convert_short(uchar);
-short __ovld __cnfn convert_short_sat(uchar);
-short __ovld __cnfn convert_short_rte(short);
-short __ovld __cnfn convert_short_sat_rte(short);
-short __ovld __cnfn convert_short_rtz(short);
-short __ovld __cnfn convert_short_sat_rtz(short);
-short __ovld __cnfn convert_short_rtp(short);
-short __ovld __cnfn convert_short_sat_rtp(short);
-short __ovld __cnfn convert_short_rtn(short);
-short __ovld __cnfn convert_short_sat_rtn(short);
-short __ovld __cnfn convert_short(short);
-short __ovld __cnfn convert_short_sat(short);
-short __ovld __cnfn convert_short_rte(ushort);
-short __ovld __cnfn convert_short_sat_rte(ushort);
-short __ovld __cnfn convert_short_rtz(ushort);
-short __ovld __cnfn convert_short_sat_rtz(ushort);
-short __ovld __cnfn convert_short_rtp(ushort);
-short __ovld __cnfn convert_short_sat_rtp(ushort);
-short __ovld __cnfn convert_short_rtn(ushort);
-short __ovld __cnfn convert_short_sat_rtn(ushort);
-short __ovld __cnfn convert_short(ushort);
-short __ovld __cnfn convert_short_sat(ushort);
-short __ovld __cnfn convert_short_rte(int);
-short __ovld __cnfn convert_short_sat_rte(int);
-short __ovld __cnfn convert_short_rtz(int);
-short __ovld __cnfn convert_short_sat_rtz(int);
-short __ovld __cnfn convert_short_rtp(int);
-short __ovld __cnfn convert_short_sat_rtp(int);
-short __ovld __cnfn convert_short_rtn(int);
-short __ovld __cnfn convert_short_sat_rtn(int);
-short __ovld __cnfn convert_short(int);
-short __ovld __cnfn convert_short_sat(int);
-short __ovld __cnfn convert_short_rte(uint);
-short __ovld __cnfn convert_short_sat_rte(uint);
-short __ovld __cnfn convert_short_rtz(uint);
-short __ovld __cnfn convert_short_sat_rtz(uint);
-short __ovld __cnfn convert_short_rtp(uint);
-short __ovld __cnfn convert_short_sat_rtp(uint);
-short __ovld __cnfn convert_short_rtn(uint);
-short __ovld __cnfn convert_short_sat_rtn(uint);
-short __ovld __cnfn convert_short(uint);
-short __ovld __cnfn convert_short_sat(uint);
-short __ovld __cnfn convert_short_rte(long);
-short __ovld __cnfn convert_short_sat_rte(long);
-short __ovld __cnfn convert_short_rtz(long);
-short __ovld __cnfn convert_short_sat_rtz(long);
-short __ovld __cnfn convert_short_rtp(long);
-short __ovld __cnfn convert_short_sat_rtp(long);
-short __ovld __cnfn convert_short_rtn(long);
-short __ovld __cnfn convert_short_sat_rtn(long);
-short __ovld __cnfn convert_short(long);
-short __ovld __cnfn convert_short_sat(long);
-short __ovld __cnfn convert_short_rte(ulong);
-short __ovld __cnfn convert_short_sat_rte(ulong);
-short __ovld __cnfn convert_short_rtz(ulong);
-short __ovld __cnfn convert_short_sat_rtz(ulong);
-short __ovld __cnfn convert_short_rtp(ulong);
-short __ovld __cnfn convert_short_sat_rtp(ulong);
-short __ovld __cnfn convert_short_rtn(ulong);
-short __ovld __cnfn convert_short_sat_rtn(ulong);
-short __ovld __cnfn convert_short(ulong);
-short __ovld __cnfn convert_short_sat(ulong);
-short __ovld __cnfn convert_short_rte(float);
-short __ovld __cnfn convert_short_sat_rte(float);
-short __ovld __cnfn convert_short_rtz(float);
-short __ovld __cnfn convert_short_sat_rtz(float);
-short __ovld __cnfn convert_short_rtp(float);
-short __ovld __cnfn convert_short_sat_rtp(float);
-short __ovld __cnfn convert_short_rtn(float);
-short __ovld __cnfn convert_short_sat_rtn(float);
-short __ovld __cnfn convert_short(float);
-short __ovld __cnfn convert_short_sat(float);
-ushort __ovld __cnfn convert_ushort_rte(char);
-ushort __ovld __cnfn convert_ushort_sat_rte(char);
-ushort __ovld __cnfn convert_ushort_rtz(char);
-ushort __ovld __cnfn convert_ushort_sat_rtz(char);
-ushort __ovld __cnfn convert_ushort_rtp(char);
-ushort __ovld __cnfn convert_ushort_sat_rtp(char);
-ushort __ovld __cnfn convert_ushort_rtn(char);
-ushort __ovld __cnfn convert_ushort_sat_rtn(char);
-ushort __ovld __cnfn convert_ushort(char);
-ushort __ovld __cnfn convert_ushort_sat(char);
-ushort __ovld __cnfn convert_ushort_rte(uchar);
-ushort __ovld __cnfn convert_ushort_sat_rte(uchar);
-ushort __ovld __cnfn convert_ushort_rtz(uchar);
-ushort __ovld __cnfn convert_ushort_sat_rtz(uchar);
-ushort __ovld __cnfn convert_ushort_rtp(uchar);
-ushort __ovld __cnfn convert_ushort_sat_rtp(uchar);
-ushort __ovld __cnfn convert_ushort_rtn(uchar);
-ushort __ovld __cnfn convert_ushort_sat_rtn(uchar);
-ushort __ovld __cnfn convert_ushort(uchar);
-ushort __ovld __cnfn convert_ushort_sat(uchar);
-ushort __ovld __cnfn convert_ushort_rte(short);
-ushort __ovld __cnfn convert_ushort_sat_rte(short);
-ushort __ovld __cnfn convert_ushort_rtz(short);
-ushort __ovld __cnfn convert_ushort_sat_rtz(short);
-ushort __ovld __cnfn convert_ushort_rtp(short);
-ushort __ovld __cnfn convert_ushort_sat_rtp(short);
-ushort __ovld __cnfn convert_ushort_rtn(short);
-ushort __ovld __cnfn convert_ushort_sat_rtn(short);
-ushort __ovld __cnfn convert_ushort(short);
-ushort __ovld __cnfn convert_ushort_sat(short);
-ushort __ovld __cnfn convert_ushort_rte(ushort);
-ushort __ovld __cnfn convert_ushort_sat_rte(ushort);
-ushort __ovld __cnfn convert_ushort_rtz(ushort);
-ushort __ovld __cnfn convert_ushort_sat_rtz(ushort);
-ushort __ovld __cnfn convert_ushort_rtp(ushort);
-ushort __ovld __cnfn convert_ushort_sat_rtp(ushort);
-ushort __ovld __cnfn convert_ushort_rtn(ushort);
-ushort __ovld __cnfn convert_ushort_sat_rtn(ushort);
-ushort __ovld __cnfn convert_ushort(ushort);
-ushort __ovld __cnfn convert_ushort_sat(ushort);
-ushort __ovld __cnfn convert_ushort_rte(int);
-ushort __ovld __cnfn convert_ushort_sat_rte(int);
-ushort __ovld __cnfn convert_ushort_rtz(int);
-ushort __ovld __cnfn convert_ushort_sat_rtz(int);
-ushort __ovld __cnfn convert_ushort_rtp(int);
-ushort __ovld __cnfn convert_ushort_sat_rtp(int);
-ushort __ovld __cnfn convert_ushort_rtn(int);
-ushort __ovld __cnfn convert_ushort_sat_rtn(int);
-ushort __ovld __cnfn convert_ushort(int);
-ushort __ovld __cnfn convert_ushort_sat(int);
-ushort __ovld __cnfn convert_ushort_rte(uint);
-ushort __ovld __cnfn convert_ushort_sat_rte(uint);
-ushort __ovld __cnfn convert_ushort_rtz(uint);
-ushort __ovld __cnfn convert_ushort_sat_rtz(uint);
-ushort __ovld __cnfn convert_ushort_rtp(uint);
-ushort __ovld __cnfn convert_ushort_sat_rtp(uint);
-ushort __ovld __cnfn convert_ushort_rtn(uint);
-ushort __ovld __cnfn convert_ushort_sat_rtn(uint);
-ushort __ovld __cnfn convert_ushort(uint);
-ushort __ovld __cnfn convert_ushort_sat(uint);
-ushort __ovld __cnfn convert_ushort_rte(long);
-ushort __ovld __cnfn convert_ushort_sat_rte(long);
-ushort __ovld __cnfn convert_ushort_rtz(long);
-ushort __ovld __cnfn convert_ushort_sat_rtz(long);
-ushort __ovld __cnfn convert_ushort_rtp(long);
-ushort __ovld __cnfn convert_ushort_sat_rtp(long);
-ushort __ovld __cnfn convert_ushort_rtn(long);
-ushort __ovld __cnfn convert_ushort_sat_rtn(long);
-ushort __ovld __cnfn convert_ushort(long);
-ushort __ovld __cnfn convert_ushort_sat(long);
-ushort __ovld __cnfn convert_ushort_rte(ulong);
-ushort __ovld __cnfn convert_ushort_sat_rte(ulong);
-ushort __ovld __cnfn convert_ushort_rtz(ulong);
-ushort __ovld __cnfn convert_ushort_sat_rtz(ulong);
-ushort __ovld __cnfn convert_ushort_rtp(ulong);
-ushort __ovld __cnfn convert_ushort_sat_rtp(ulong);
-ushort __ovld __cnfn convert_ushort_rtn(ulong);
-ushort __ovld __cnfn convert_ushort_sat_rtn(ulong);
-ushort __ovld __cnfn convert_ushort(ulong);
-ushort __ovld __cnfn convert_ushort_sat(ulong);
-ushort __ovld __cnfn convert_ushort_rte(float);
-ushort __ovld __cnfn convert_ushort_sat_rte(float);
-ushort __ovld __cnfn convert_ushort_rtz(float);
-ushort __ovld __cnfn convert_ushort_sat_rtz(float);
-ushort __ovld __cnfn convert_ushort_rtp(float);
-ushort __ovld __cnfn convert_ushort_sat_rtp(float);
-ushort __ovld __cnfn convert_ushort_rtn(float);
-ushort __ovld __cnfn convert_ushort_sat_rtn(float);
-ushort __ovld __cnfn convert_ushort(float);
-ushort __ovld __cnfn convert_ushort_sat(float);
-int __ovld __cnfn convert_int_rte(char);
-int __ovld __cnfn convert_int_sat_rte(char);
-int __ovld __cnfn convert_int_rtz(char);
-int __ovld __cnfn convert_int_sat_rtz(char);
-int __ovld __cnfn convert_int_rtp(char);
-int __ovld __cnfn convert_int_sat_rtp(char);
-int __ovld __cnfn convert_int_rtn(char);
-int __ovld __cnfn convert_int_sat_rtn(char);
-int __ovld __cnfn convert_int(char);
-int __ovld __cnfn convert_int_sat(char);
-int __ovld __cnfn convert_int_rte(uchar);
-int __ovld __cnfn convert_int_sat_rte(uchar);
-int __ovld __cnfn convert_int_rtz(uchar);
-int __ovld __cnfn convert_int_sat_rtz(uchar);
-int __ovld __cnfn convert_int_rtp(uchar);
-int __ovld __cnfn convert_int_sat_rtp(uchar);
-int __ovld __cnfn convert_int_rtn(uchar);
-int __ovld __cnfn convert_int_sat_rtn(uchar);
-int __ovld __cnfn convert_int(uchar);
-int __ovld __cnfn convert_int_sat(uchar);
-int __ovld __cnfn convert_int_rte(short);
-int __ovld __cnfn convert_int_sat_rte(short);
-int __ovld __cnfn convert_int_rtz(short);
-int __ovld __cnfn convert_int_sat_rtz(short);
-int __ovld __cnfn convert_int_rtp(short);
-int __ovld __cnfn convert_int_sat_rtp(short);
-int __ovld __cnfn convert_int_rtn(short);
-int __ovld __cnfn convert_int_sat_rtn(short);
-int __ovld __cnfn convert_int(short);
-int __ovld __cnfn convert_int_sat(short);
-int __ovld __cnfn convert_int_rte(ushort);
-int __ovld __cnfn convert_int_sat_rte(ushort);
-int __ovld __cnfn convert_int_rtz(ushort);
-int __ovld __cnfn convert_int_sat_rtz(ushort);
-int __ovld __cnfn convert_int_rtp(ushort);
-int __ovld __cnfn convert_int_sat_rtp(ushort);
-int __ovld __cnfn convert_int_rtn(ushort);
-int __ovld __cnfn convert_int_sat_rtn(ushort);
-int __ovld __cnfn convert_int(ushort);
-int __ovld __cnfn convert_int_sat(ushort);
-int __ovld __cnfn convert_int_rte(int);
-int __ovld __cnfn convert_int_sat_rte(int);
-int __ovld __cnfn convert_int_rtz(int);
-int __ovld __cnfn convert_int_sat_rtz(int);
-int __ovld __cnfn convert_int_rtp(int);
-int __ovld __cnfn convert_int_sat_rtp(int);
-int __ovld __cnfn convert_int_rtn(int);
-int __ovld __cnfn convert_int_sat_rtn(int);
-int __ovld __cnfn convert_int(int);
-int __ovld __cnfn convert_int_sat(int);
-int __ovld __cnfn convert_int_rte(uint);
-int __ovld __cnfn convert_int_sat_rte(uint);
-int __ovld __cnfn convert_int_rtz(uint);
-int __ovld __cnfn convert_int_sat_rtz(uint);
-int __ovld __cnfn convert_int_rtp(uint);
-int __ovld __cnfn convert_int_sat_rtp(uint);
-int __ovld __cnfn convert_int_rtn(uint);
-int __ovld __cnfn convert_int_sat_rtn(uint);
-int __ovld __cnfn convert_int(uint);
-int __ovld __cnfn convert_int_sat(uint);
-int __ovld __cnfn convert_int_rte(long);
-int __ovld __cnfn convert_int_sat_rte(long);
-int __ovld __cnfn convert_int_rtz(long);
-int __ovld __cnfn convert_int_sat_rtz(long);
-int __ovld __cnfn convert_int_rtp(long);
-int __ovld __cnfn convert_int_sat_rtp(long);
-int __ovld __cnfn convert_int_rtn(long);
-int __ovld __cnfn convert_int_sat_rtn(long);
-int __ovld __cnfn convert_int(long);
-int __ovld __cnfn convert_int_sat(long);
-int __ovld __cnfn convert_int_rte(ulong);
-int __ovld __cnfn convert_int_sat_rte(ulong);
-int __ovld __cnfn convert_int_rtz(ulong);
-int __ovld __cnfn convert_int_sat_rtz(ulong);
-int __ovld __cnfn convert_int_rtp(ulong);
-int __ovld __cnfn convert_int_sat_rtp(ulong);
-int __ovld __cnfn convert_int_rtn(ulong);
-int __ovld __cnfn convert_int_sat_rtn(ulong);
-int __ovld __cnfn convert_int(ulong);
-int __ovld __cnfn convert_int_sat(ulong);
-int __ovld __cnfn convert_int_rte(float);
-int __ovld __cnfn convert_int_sat_rte(float);
-int __ovld __cnfn convert_int_rtz(float);
-int __ovld __cnfn convert_int_sat_rtz(float);
-int __ovld __cnfn convert_int_rtp(float);
-int __ovld __cnfn convert_int_sat_rtp(float);
-int __ovld __cnfn convert_int_rtn(float);
-int __ovld __cnfn convert_int_sat_rtn(float);
-int __ovld __cnfn convert_int(float);
-int __ovld __cnfn convert_int_sat(float);
-uint __ovld __cnfn convert_uint_rte(char);
-uint __ovld __cnfn convert_uint_sat_rte(char);
-uint __ovld __cnfn convert_uint_rtz(char);
-uint __ovld __cnfn convert_uint_sat_rtz(char);
-uint __ovld __cnfn convert_uint_rtp(char);
-uint __ovld __cnfn convert_uint_sat_rtp(char);
-uint __ovld __cnfn convert_uint_rtn(char);
-uint __ovld __cnfn convert_uint_sat_rtn(char);
-uint __ovld __cnfn convert_uint(char);
-uint __ovld __cnfn convert_uint_sat(char);
-uint __ovld __cnfn convert_uint_rte(uchar);
-uint __ovld __cnfn convert_uint_sat_rte(uchar);
-uint __ovld __cnfn convert_uint_rtz(uchar);
-uint __ovld __cnfn convert_uint_sat_rtz(uchar);
-uint __ovld __cnfn convert_uint_rtp(uchar);
-uint __ovld __cnfn convert_uint_sat_rtp(uchar);
-uint __ovld __cnfn convert_uint_rtn(uchar);
-uint __ovld __cnfn convert_uint_sat_rtn(uchar);
-uint __ovld __cnfn convert_uint(uchar);
-uint __ovld __cnfn convert_uint_sat(uchar);
-uint __ovld __cnfn convert_uint_rte(short);
-uint __ovld __cnfn convert_uint_sat_rte(short);
-uint __ovld __cnfn convert_uint_rtz(short);
-uint __ovld __cnfn convert_uint_sat_rtz(short);
-uint __ovld __cnfn convert_uint_rtp(short);
-uint __ovld __cnfn convert_uint_sat_rtp(short);
-uint __ovld __cnfn convert_uint_rtn(short);
-uint __ovld __cnfn convert_uint_sat_rtn(short);
-uint __ovld __cnfn convert_uint(short);
-uint __ovld __cnfn convert_uint_sat(short);
-uint __ovld __cnfn convert_uint_rte(ushort);
-uint __ovld __cnfn convert_uint_sat_rte(ushort);
-uint __ovld __cnfn convert_uint_rtz(ushort);
-uint __ovld __cnfn convert_uint_sat_rtz(ushort);
-uint __ovld __cnfn convert_uint_rtp(ushort);
-uint __ovld __cnfn convert_uint_sat_rtp(ushort);
-uint __ovld __cnfn convert_uint_rtn(ushort);
-uint __ovld __cnfn convert_uint_sat_rtn(ushort);
-uint __ovld __cnfn convert_uint(ushort);
-uint __ovld __cnfn convert_uint_sat(ushort);
-uint __ovld __cnfn convert_uint_rte(int);
-uint __ovld __cnfn convert_uint_sat_rte(int);
-uint __ovld __cnfn convert_uint_rtz(int);
-uint __ovld __cnfn convert_uint_sat_rtz(int);
-uint __ovld __cnfn convert_uint_rtp(int);
-uint __ovld __cnfn convert_uint_sat_rtp(int);
-uint __ovld __cnfn convert_uint_rtn(int);
-uint __ovld __cnfn convert_uint_sat_rtn(int);
-uint __ovld __cnfn convert_uint(int);
-uint __ovld __cnfn convert_uint_sat(int);
-uint __ovld __cnfn convert_uint_rte(uint);
-uint __ovld __cnfn convert_uint_sat_rte(uint);
-uint __ovld __cnfn convert_uint_rtz(uint);
-uint __ovld __cnfn convert_uint_sat_rtz(uint);
-uint __ovld __cnfn convert_uint_rtp(uint);
-uint __ovld __cnfn convert_uint_sat_rtp(uint);
-uint __ovld __cnfn convert_uint_rtn(uint);
-uint __ovld __cnfn convert_uint_sat_rtn(uint);
-uint __ovld __cnfn convert_uint(uint);
-uint __ovld __cnfn convert_uint_sat(uint);
-uint __ovld __cnfn convert_uint_rte(long);
-uint __ovld __cnfn convert_uint_sat_rte(long);
-uint __ovld __cnfn convert_uint_rtz(long);
-uint __ovld __cnfn convert_uint_sat_rtz(long);
-uint __ovld __cnfn convert_uint_rtp(long);
-uint __ovld __cnfn convert_uint_sat_rtp(long);
-uint __ovld __cnfn convert_uint_rtn(long);
-uint __ovld __cnfn convert_uint_sat_rtn(long);
-uint __ovld __cnfn convert_uint(long);
-uint __ovld __cnfn convert_uint_sat(long);
-uint __ovld __cnfn convert_uint_rte(ulong);
-uint __ovld __cnfn convert_uint_sat_rte(ulong);
-uint __ovld __cnfn convert_uint_rtz(ulong);
-uint __ovld __cnfn convert_uint_sat_rtz(ulong);
-uint __ovld __cnfn convert_uint_rtp(ulong);
-uint __ovld __cnfn convert_uint_sat_rtp(ulong);
-uint __ovld __cnfn convert_uint_rtn(ulong);
-uint __ovld __cnfn convert_uint_sat_rtn(ulong);
-uint __ovld __cnfn convert_uint(ulong);
-uint __ovld __cnfn convert_uint_sat(ulong);
-uint __ovld __cnfn convert_uint_rte(float);
-uint __ovld __cnfn convert_uint_sat_rte(float);
-uint __ovld __cnfn convert_uint_rtz(float);
-uint __ovld __cnfn convert_uint_sat_rtz(float);
-uint __ovld __cnfn convert_uint_rtp(float);
-uint __ovld __cnfn convert_uint_sat_rtp(float);
-uint __ovld __cnfn convert_uint_rtn(float);
-uint __ovld __cnfn convert_uint_sat_rtn(float);
-uint __ovld __cnfn convert_uint(float);
-uint __ovld __cnfn convert_uint_sat(float);
-long __ovld __cnfn convert_long_rte(char);
-long __ovld __cnfn convert_long_sat_rte(char);
-long __ovld __cnfn convert_long_rtz(char);
-long __ovld __cnfn convert_long_sat_rtz(char);
-long __ovld __cnfn convert_long_rtp(char);
-long __ovld __cnfn convert_long_sat_rtp(char);
-long __ovld __cnfn convert_long_rtn(char);
-long __ovld __cnfn convert_long_sat_rtn(char);
-long __ovld __cnfn convert_long(char);
-long __ovld __cnfn convert_long_sat(char);
-long __ovld __cnfn convert_long_rte(uchar);
-long __ovld __cnfn convert_long_sat_rte(uchar);
-long __ovld __cnfn convert_long_rtz(uchar);
-long __ovld __cnfn convert_long_sat_rtz(uchar);
-long __ovld __cnfn convert_long_rtp(uchar);
-long __ovld __cnfn convert_long_sat_rtp(uchar);
-long __ovld __cnfn convert_long_rtn(uchar);
-long __ovld __cnfn convert_long_sat_rtn(uchar);
-long __ovld __cnfn convert_long(uchar);
-long __ovld __cnfn convert_long_sat(uchar);
-long __ovld __cnfn convert_long_rte(short);
-long __ovld __cnfn convert_long_sat_rte(short);
-long __ovld __cnfn convert_long_rtz(short);
-long __ovld __cnfn convert_long_sat_rtz(short);
-long __ovld __cnfn convert_long_rtp(short);
-long __ovld __cnfn convert_long_sat_rtp(short);
-long __ovld __cnfn convert_long_rtn(short);
-long __ovld __cnfn convert_long_sat_rtn(short);
-long __ovld __cnfn convert_long(short);
-long __ovld __cnfn convert_long_sat(short);
-long __ovld __cnfn convert_long_rte(ushort);
-long __ovld __cnfn convert_long_sat_rte(ushort);
-long __ovld __cnfn convert_long_rtz(ushort);
-long __ovld __cnfn convert_long_sat_rtz(ushort);
-long __ovld __cnfn convert_long_rtp(ushort);
-long __ovld __cnfn convert_long_sat_rtp(ushort);
-long __ovld __cnfn convert_long_rtn(ushort);
-long __ovld __cnfn convert_long_sat_rtn(ushort);
-long __ovld __cnfn convert_long(ushort);
-long __ovld __cnfn convert_long_sat(ushort);
-long __ovld __cnfn convert_long_rte(int);
-long __ovld __cnfn convert_long_sat_rte(int);
-long __ovld __cnfn convert_long_rtz(int);
-long __ovld __cnfn convert_long_sat_rtz(int);
-long __ovld __cnfn convert_long_rtp(int);
-long __ovld __cnfn convert_long_sat_rtp(int);
-long __ovld __cnfn convert_long_rtn(int);
-long __ovld __cnfn convert_long_sat_rtn(int);
-long __ovld __cnfn convert_long(int);
-long __ovld __cnfn convert_long_sat(int);
-long __ovld __cnfn convert_long_rte(uint);
-long __ovld __cnfn convert_long_sat_rte(uint);
-long __ovld __cnfn convert_long_rtz(uint);
-long __ovld __cnfn convert_long_sat_rtz(uint);
-long __ovld __cnfn convert_long_rtp(uint);
-long __ovld __cnfn convert_long_sat_rtp(uint);
-long __ovld __cnfn convert_long_rtn(uint);
-long __ovld __cnfn convert_long_sat_rtn(uint);
-long __ovld __cnfn convert_long(uint);
-long __ovld __cnfn convert_long_sat(uint);
-long __ovld __cnfn convert_long_rte(long);
-long __ovld __cnfn convert_long_sat_rte(long);
-long __ovld __cnfn convert_long_rtz(long);
-long __ovld __cnfn convert_long_sat_rtz(long);
-long __ovld __cnfn convert_long_rtp(long);
-long __ovld __cnfn convert_long_sat_rtp(long);
-long __ovld __cnfn convert_long_rtn(long);
-long __ovld __cnfn convert_long_sat_rtn(long);
-long __ovld __cnfn convert_long(long);
-long __ovld __cnfn convert_long_sat(long);
-long __ovld __cnfn convert_long_rte(ulong);
-long __ovld __cnfn convert_long_sat_rte(ulong);
-long __ovld __cnfn convert_long_rtz(ulong);
-long __ovld __cnfn convert_long_sat_rtz(ulong);
-long __ovld __cnfn convert_long_rtp(ulong);
-long __ovld __cnfn convert_long_sat_rtp(ulong);
-long __ovld __cnfn convert_long_rtn(ulong);
-long __ovld __cnfn convert_long_sat_rtn(ulong);
-long __ovld __cnfn convert_long(ulong);
-long __ovld __cnfn convert_long_sat(ulong);
-long __ovld __cnfn convert_long_rte(float);
-long __ovld __cnfn convert_long_sat_rte(float);
-long __ovld __cnfn convert_long_rtz(float);
-long __ovld __cnfn convert_long_sat_rtz(float);
-long __ovld __cnfn convert_long_rtp(float);
-long __ovld __cnfn convert_long_sat_rtp(float);
-long __ovld __cnfn convert_long_rtn(float);
-long __ovld __cnfn convert_long_sat_rtn(float);
-long __ovld __cnfn convert_long(float);
-long __ovld __cnfn convert_long_sat(float);
-ulong __ovld __cnfn convert_ulong_rte(char);
-ulong __ovld __cnfn convert_ulong_sat_rte(char);
-ulong __ovld __cnfn convert_ulong_rtz(char);
-ulong __ovld __cnfn convert_ulong_sat_rtz(char);
-ulong __ovld __cnfn convert_ulong_rtp(char);
-ulong __ovld __cnfn convert_ulong_sat_rtp(char);
-ulong __ovld __cnfn convert_ulong_rtn(char);
-ulong __ovld __cnfn convert_ulong_sat_rtn(char);
-ulong __ovld __cnfn convert_ulong(char);
-ulong __ovld __cnfn convert_ulong_sat(char);
-ulong __ovld __cnfn convert_ulong_rte(uchar);
-ulong __ovld __cnfn convert_ulong_sat_rte(uchar);
-ulong __ovld __cnfn convert_ulong_rtz(uchar);
-ulong __ovld __cnfn convert_ulong_sat_rtz(uchar);
-ulong __ovld __cnfn convert_ulong_rtp(uchar);
-ulong __ovld __cnfn convert_ulong_sat_rtp(uchar);
-ulong __ovld __cnfn convert_ulong_rtn(uchar);
-ulong __ovld __cnfn convert_ulong_sat_rtn(uchar);
-ulong __ovld __cnfn convert_ulong(uchar);
-ulong __ovld __cnfn convert_ulong_sat(uchar);
-ulong __ovld __cnfn convert_ulong_rte(short);
-ulong __ovld __cnfn convert_ulong_sat_rte(short);
-ulong __ovld __cnfn convert_ulong_rtz(short);
-ulong __ovld __cnfn convert_ulong_sat_rtz(short);
-ulong __ovld __cnfn convert_ulong_rtp(short);
-ulong __ovld __cnfn convert_ulong_sat_rtp(short);
-ulong __ovld __cnfn convert_ulong_rtn(short);
-ulong __ovld __cnfn convert_ulong_sat_rtn(short);
-ulong __ovld __cnfn convert_ulong(short);
-ulong __ovld __cnfn convert_ulong_sat(short);
-ulong __ovld __cnfn convert_ulong_rte(ushort);
-ulong __ovld __cnfn convert_ulong_sat_rte(ushort);
-ulong __ovld __cnfn convert_ulong_rtz(ushort);
-ulong __ovld __cnfn convert_ulong_sat_rtz(ushort);
-ulong __ovld __cnfn convert_ulong_rtp(ushort);
-ulong __ovld __cnfn convert_ulong_sat_rtp(ushort);
-ulong __ovld __cnfn convert_ulong_rtn(ushort);
-ulong __ovld __cnfn convert_ulong_sat_rtn(ushort);
-ulong __ovld __cnfn convert_ulong(ushort);
-ulong __ovld __cnfn convert_ulong_sat(ushort);
-ulong __ovld __cnfn convert_ulong_rte(int);
-ulong __ovld __cnfn convert_ulong_sat_rte(int);
-ulong __ovld __cnfn convert_ulong_rtz(int);
-ulong __ovld __cnfn convert_ulong_sat_rtz(int);
-ulong __ovld __cnfn convert_ulong_rtp(int);
-ulong __ovld __cnfn convert_ulong_sat_rtp(int);
-ulong __ovld __cnfn convert_ulong_rtn(int);
-ulong __ovld __cnfn convert_ulong_sat_rtn(int);
-ulong __ovld __cnfn convert_ulong(int);
-ulong __ovld __cnfn convert_ulong_sat(int);
-ulong __ovld __cnfn convert_ulong_rte(uint);
-ulong __ovld __cnfn convert_ulong_sat_rte(uint);
-ulong __ovld __cnfn convert_ulong_rtz(uint);
-ulong __ovld __cnfn convert_ulong_sat_rtz(uint);
-ulong __ovld __cnfn convert_ulong_rtp(uint);
-ulong __ovld __cnfn convert_ulong_sat_rtp(uint);
-ulong __ovld __cnfn convert_ulong_rtn(uint);
-ulong __ovld __cnfn convert_ulong_sat_rtn(uint);
-ulong __ovld __cnfn convert_ulong(uint);
-ulong __ovld __cnfn convert_ulong_sat(uint);
-ulong __ovld __cnfn convert_ulong_rte(long);
-ulong __ovld __cnfn convert_ulong_sat_rte(long);
-ulong __ovld __cnfn convert_ulong_rtz(long);
-ulong __ovld __cnfn convert_ulong_sat_rtz(long);
-ulong __ovld __cnfn convert_ulong_rtp(long);
-ulong __ovld __cnfn convert_ulong_sat_rtp(long);
-ulong __ovld __cnfn convert_ulong_rtn(long);
-ulong __ovld __cnfn convert_ulong_sat_rtn(long);
-ulong __ovld __cnfn convert_ulong(long);
-ulong __ovld __cnfn convert_ulong_sat(long);
-ulong __ovld __cnfn convert_ulong_rte(ulong);
-ulong __ovld __cnfn convert_ulong_sat_rte(ulong);
-ulong __ovld __cnfn convert_ulong_rtz(ulong);
-ulong __ovld __cnfn convert_ulong_sat_rtz(ulong);
-ulong __ovld __cnfn convert_ulong_rtp(ulong);
-ulong __ovld __cnfn convert_ulong_sat_rtp(ulong);
-ulong __ovld __cnfn convert_ulong_rtn(ulong);
-ulong __ovld __cnfn convert_ulong_sat_rtn(ulong);
-ulong __ovld __cnfn convert_ulong(ulong);
-ulong __ovld __cnfn convert_ulong_sat(ulong);
-ulong __ovld __cnfn convert_ulong_rte(float);
-ulong __ovld __cnfn convert_ulong_sat_rte(float);
-ulong __ovld __cnfn convert_ulong_rtz(float);
-ulong __ovld __cnfn convert_ulong_sat_rtz(float);
-ulong __ovld __cnfn convert_ulong_rtp(float);
-ulong __ovld __cnfn convert_ulong_sat_rtp(float);
-ulong __ovld __cnfn convert_ulong_rtn(float);
-ulong __ovld __cnfn convert_ulong_sat_rtn(float);
-ulong __ovld __cnfn convert_ulong(float);
-ulong __ovld __cnfn convert_ulong_sat(float);
-float __ovld __cnfn convert_float_rte(char);
-float __ovld __cnfn convert_float_rtz(char);
-float __ovld __cnfn convert_float_rtp(char);
-float __ovld __cnfn convert_float_rtn(char);
-float __ovld __cnfn convert_float(char);
-float __ovld __cnfn convert_float_rte(uchar);
-float __ovld __cnfn convert_float_rtz(uchar);
-float __ovld __cnfn convert_float_rtp(uchar);
-float __ovld __cnfn convert_float_rtn(uchar);
-float __ovld __cnfn convert_float(uchar);
-float __ovld __cnfn convert_float_rte(short);
-float __ovld __cnfn convert_float_rtz(short);
-float __ovld __cnfn convert_float_rtp(short);
-float __ovld __cnfn convert_float_rtn(short);
-float __ovld __cnfn convert_float(short);
-float __ovld __cnfn convert_float_rte(ushort);
-float __ovld __cnfn convert_float_rtz(ushort);
-float __ovld __cnfn convert_float_rtp(ushort);
-float __ovld __cnfn convert_float_rtn(ushort);
-float __ovld __cnfn convert_float(ushort);
-float __ovld __cnfn convert_float_rte(int);
-float __ovld __cnfn convert_float_rtz(int);
-float __ovld __cnfn convert_float_rtp(int);
-float __ovld __cnfn convert_float_rtn(int);
-float __ovld __cnfn convert_float(int);
-float __ovld __cnfn convert_float_rte(uint);
-float __ovld __cnfn convert_float_rtz(uint);
-float __ovld __cnfn convert_float_rtp(uint);
-float __ovld __cnfn convert_float_rtn(uint);
-float __ovld __cnfn convert_float(uint);
-float __ovld __cnfn convert_float_rte(long);
-float __ovld __cnfn convert_float_rtz(long);
-float __ovld __cnfn convert_float_rtp(long);
-float __ovld __cnfn convert_float_rtn(long);
-float __ovld __cnfn convert_float(long);
-float __ovld __cnfn convert_float_rte(ulong);
-float __ovld __cnfn convert_float_rtz(ulong);
-float __ovld __cnfn convert_float_rtp(ulong);
-float __ovld __cnfn convert_float_rtn(ulong);
-float __ovld __cnfn convert_float(ulong);
-float __ovld __cnfn convert_float_rte(float);
-float __ovld __cnfn convert_float_rtz(float);
-float __ovld __cnfn convert_float_rtp(float);
-float __ovld __cnfn convert_float_rtn(float);
-float __ovld __cnfn convert_float(float);
-char2 __ovld __cnfn convert_char2_rte(char2);
-char2 __ovld __cnfn convert_char2_sat_rte(char2);
-char2 __ovld __cnfn convert_char2_rtz(char2);
-char2 __ovld __cnfn convert_char2_sat_rtz(char2);
-char2 __ovld __cnfn convert_char2_rtp(char2);
-char2 __ovld __cnfn convert_char2_sat_rtp(char2);
-char2 __ovld __cnfn convert_char2_rtn(char2);
-char2 __ovld __cnfn convert_char2_sat_rtn(char2);
-char2 __ovld __cnfn convert_char2(char2);
-char2 __ovld __cnfn convert_char2_sat(char2);
-char2 __ovld __cnfn convert_char2_rte(uchar2);
-char2 __ovld __cnfn convert_char2_sat_rte(uchar2);
-char2 __ovld __cnfn convert_char2_rtz(uchar2);
-char2 __ovld __cnfn convert_char2_sat_rtz(uchar2);
-char2 __ovld __cnfn convert_char2_rtp(uchar2);
-char2 __ovld __cnfn convert_char2_sat_rtp(uchar2);
-char2 __ovld __cnfn convert_char2_rtn(uchar2);
-char2 __ovld __cnfn convert_char2_sat_rtn(uchar2);
-char2 __ovld __cnfn convert_char2(uchar2);
-char2 __ovld __cnfn convert_char2_sat(uchar2);
-char2 __ovld __cnfn convert_char2_rte(short2);
-char2 __ovld __cnfn convert_char2_sat_rte(short2);
-char2 __ovld __cnfn convert_char2_rtz(short2);
-char2 __ovld __cnfn convert_char2_sat_rtz(short2);
-char2 __ovld __cnfn convert_char2_rtp(short2);
-char2 __ovld __cnfn convert_char2_sat_rtp(short2);
-char2 __ovld __cnfn convert_char2_rtn(short2);
-char2 __ovld __cnfn convert_char2_sat_rtn(short2);
-char2 __ovld __cnfn convert_char2(short2);
-char2 __ovld __cnfn convert_char2_sat(short2);
-char2 __ovld __cnfn convert_char2_rte(ushort2);
-char2 __ovld __cnfn convert_char2_sat_rte(ushort2);
-char2 __ovld __cnfn convert_char2_rtz(ushort2);
-char2 __ovld __cnfn convert_char2_sat_rtz(ushort2);
-char2 __ovld __cnfn convert_char2_rtp(ushort2);
-char2 __ovld __cnfn convert_char2_sat_rtp(ushort2);
-char2 __ovld __cnfn convert_char2_rtn(ushort2);
-char2 __ovld __cnfn convert_char2_sat_rtn(ushort2);
-char2 __ovld __cnfn convert_char2(ushort2);
-char2 __ovld __cnfn convert_char2_sat(ushort2);
-char2 __ovld __cnfn convert_char2_rte(int2);
-char2 __ovld __cnfn convert_char2_sat_rte(int2);
-char2 __ovld __cnfn convert_char2_rtz(int2);
-char2 __ovld __cnfn convert_char2_sat_rtz(int2);
-char2 __ovld __cnfn convert_char2_rtp(int2);
-char2 __ovld __cnfn convert_char2_sat_rtp(int2);
-char2 __ovld __cnfn convert_char2_rtn(int2);
-char2 __ovld __cnfn convert_char2_sat_rtn(int2);
-char2 __ovld __cnfn convert_char2(int2);
-char2 __ovld __cnfn convert_char2_sat(int2);
-char2 __ovld __cnfn convert_char2_rte(uint2);
-char2 __ovld __cnfn convert_char2_sat_rte(uint2);
-char2 __ovld __cnfn convert_char2_rtz(uint2);
-char2 __ovld __cnfn convert_char2_sat_rtz(uint2);
-char2 __ovld __cnfn convert_char2_rtp(uint2);
-char2 __ovld __cnfn convert_char2_sat_rtp(uint2);
-char2 __ovld __cnfn convert_char2_rtn(uint2);
-char2 __ovld __cnfn convert_char2_sat_rtn(uint2);
-char2 __ovld __cnfn convert_char2(uint2);
-char2 __ovld __cnfn convert_char2_sat(uint2);
-char2 __ovld __cnfn convert_char2_rte(long2);
-char2 __ovld __cnfn convert_char2_sat_rte(long2);
-char2 __ovld __cnfn convert_char2_rtz(long2);
-char2 __ovld __cnfn convert_char2_sat_rtz(long2);
-char2 __ovld __cnfn convert_char2_rtp(long2);
-char2 __ovld __cnfn convert_char2_sat_rtp(long2);
-char2 __ovld __cnfn convert_char2_rtn(long2);
-char2 __ovld __cnfn convert_char2_sat_rtn(long2);
-char2 __ovld __cnfn convert_char2(long2);
-char2 __ovld __cnfn convert_char2_sat(long2);
-char2 __ovld __cnfn convert_char2_rte(ulong2);
-char2 __ovld __cnfn convert_char2_sat_rte(ulong2);
-char2 __ovld __cnfn convert_char2_rtz(ulong2);
-char2 __ovld __cnfn convert_char2_sat_rtz(ulong2);
-char2 __ovld __cnfn convert_char2_rtp(ulong2);
-char2 __ovld __cnfn convert_char2_sat_rtp(ulong2);
-char2 __ovld __cnfn convert_char2_rtn(ulong2);
-char2 __ovld __cnfn convert_char2_sat_rtn(ulong2);
-char2 __ovld __cnfn convert_char2(ulong2);
-char2 __ovld __cnfn convert_char2_sat(ulong2);
-char2 __ovld __cnfn convert_char2_rte(float2);
-char2 __ovld __cnfn convert_char2_sat_rte(float2);
-char2 __ovld __cnfn convert_char2_rtz(float2);
-char2 __ovld __cnfn convert_char2_sat_rtz(float2);
-char2 __ovld __cnfn convert_char2_rtp(float2);
-char2 __ovld __cnfn convert_char2_sat_rtp(float2);
-char2 __ovld __cnfn convert_char2_rtn(float2);
-char2 __ovld __cnfn convert_char2_sat_rtn(float2);
-char2 __ovld __cnfn convert_char2(float2);
-char2 __ovld __cnfn convert_char2_sat(float2);
-uchar2 __ovld __cnfn convert_uchar2_rte(char2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(char2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(char2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(char2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(char2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(char2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(char2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(char2);
-uchar2 __ovld __cnfn convert_uchar2(char2);
-uchar2 __ovld __cnfn convert_uchar2_sat(char2);
-uchar2 __ovld __cnfn convert_uchar2_rte(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(uchar2);
-uchar2 __ovld __cnfn convert_uchar2(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_sat(uchar2);
-uchar2 __ovld __cnfn convert_uchar2_rte(short2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(short2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(short2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(short2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(short2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(short2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(short2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(short2);
-uchar2 __ovld __cnfn convert_uchar2(short2);
-uchar2 __ovld __cnfn convert_uchar2_sat(short2);
-uchar2 __ovld __cnfn convert_uchar2_rte(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(ushort2);
-uchar2 __ovld __cnfn convert_uchar2(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_sat(ushort2);
-uchar2 __ovld __cnfn convert_uchar2_rte(int2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(int2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(int2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(int2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(int2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(int2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(int2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(int2);
-uchar2 __ovld __cnfn convert_uchar2(int2);
-uchar2 __ovld __cnfn convert_uchar2_sat(int2);
-uchar2 __ovld __cnfn convert_uchar2_rte(uint2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(uint2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(uint2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(uint2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(uint2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(uint2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(uint2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(uint2);
-uchar2 __ovld __cnfn convert_uchar2(uint2);
-uchar2 __ovld __cnfn convert_uchar2_sat(uint2);
-uchar2 __ovld __cnfn convert_uchar2_rte(long2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(long2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(long2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(long2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(long2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(long2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(long2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(long2);
-uchar2 __ovld __cnfn convert_uchar2(long2);
-uchar2 __ovld __cnfn convert_uchar2_sat(long2);
-uchar2 __ovld __cnfn convert_uchar2_rte(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(ulong2);
-uchar2 __ovld __cnfn convert_uchar2(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_sat(ulong2);
-uchar2 __ovld __cnfn convert_uchar2_rte(float2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(float2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(float2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(float2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(float2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(float2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(float2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(float2);
-uchar2 __ovld __cnfn convert_uchar2(float2);
-uchar2 __ovld __cnfn convert_uchar2_sat(float2);
-short2 __ovld __cnfn convert_short2_rte(char2);
-short2 __ovld __cnfn convert_short2_sat_rte(char2);
-short2 __ovld __cnfn convert_short2_rtz(char2);
-short2 __ovld __cnfn convert_short2_sat_rtz(char2);
-short2 __ovld __cnfn convert_short2_rtp(char2);
-short2 __ovld __cnfn convert_short2_sat_rtp(char2);
-short2 __ovld __cnfn convert_short2_rtn(char2);
-short2 __ovld __cnfn convert_short2_sat_rtn(char2);
-short2 __ovld __cnfn convert_short2(char2);
-short2 __ovld __cnfn convert_short2_sat(char2);
-short2 __ovld __cnfn convert_short2_rte(uchar2);
-short2 __ovld __cnfn convert_short2_sat_rte(uchar2);
-short2 __ovld __cnfn convert_short2_rtz(uchar2);
-short2 __ovld __cnfn convert_short2_sat_rtz(uchar2);
-short2 __ovld __cnfn convert_short2_rtp(uchar2);
-short2 __ovld __cnfn convert_short2_sat_rtp(uchar2);
-short2 __ovld __cnfn convert_short2_rtn(uchar2);
-short2 __ovld __cnfn convert_short2_sat_rtn(uchar2);
-short2 __ovld __cnfn convert_short2(uchar2);
-short2 __ovld __cnfn convert_short2_sat(uchar2);
-short2 __ovld __cnfn convert_short2_rte(short2);
-short2 __ovld __cnfn convert_short2_sat_rte(short2);
-short2 __ovld __cnfn convert_short2_rtz(short2);
-short2 __ovld __cnfn convert_short2_sat_rtz(short2);
-short2 __ovld __cnfn convert_short2_rtp(short2);
-short2 __ovld __cnfn convert_short2_sat_rtp(short2);
-short2 __ovld __cnfn convert_short2_rtn(short2);
-short2 __ovld __cnfn convert_short2_sat_rtn(short2);
-short2 __ovld __cnfn convert_short2(short2);
-short2 __ovld __cnfn convert_short2_sat(short2);
-short2 __ovld __cnfn convert_short2_rte(ushort2);
-short2 __ovld __cnfn convert_short2_sat_rte(ushort2);
-short2 __ovld __cnfn convert_short2_rtz(ushort2);
-short2 __ovld __cnfn convert_short2_sat_rtz(ushort2);
-short2 __ovld __cnfn convert_short2_rtp(ushort2);
-short2 __ovld __cnfn convert_short2_sat_rtp(ushort2);
-short2 __ovld __cnfn convert_short2_rtn(ushort2);
-short2 __ovld __cnfn convert_short2_sat_rtn(ushort2);
-short2 __ovld __cnfn convert_short2(ushort2);
-short2 __ovld __cnfn convert_short2_sat(ushort2);
-short2 __ovld __cnfn convert_short2_rte(int2);
-short2 __ovld __cnfn convert_short2_sat_rte(int2);
-short2 __ovld __cnfn convert_short2_rtz(int2);
-short2 __ovld __cnfn convert_short2_sat_rtz(int2);
-short2 __ovld __cnfn convert_short2_rtp(int2);
-short2 __ovld __cnfn convert_short2_sat_rtp(int2);
-short2 __ovld __cnfn convert_short2_rtn(int2);
-short2 __ovld __cnfn convert_short2_sat_rtn(int2);
-short2 __ovld __cnfn convert_short2(int2);
-short2 __ovld __cnfn convert_short2_sat(int2);
-short2 __ovld __cnfn convert_short2_rte(uint2);
-short2 __ovld __cnfn convert_short2_sat_rte(uint2);
-short2 __ovld __cnfn convert_short2_rtz(uint2);
-short2 __ovld __cnfn convert_short2_sat_rtz(uint2);
-short2 __ovld __cnfn convert_short2_rtp(uint2);
-short2 __ovld __cnfn convert_short2_sat_rtp(uint2);
-short2 __ovld __cnfn convert_short2_rtn(uint2);
-short2 __ovld __cnfn convert_short2_sat_rtn(uint2);
-short2 __ovld __cnfn convert_short2(uint2);
-short2 __ovld __cnfn convert_short2_sat(uint2);
-short2 __ovld __cnfn convert_short2_rte(long2);
-short2 __ovld __cnfn convert_short2_sat_rte(long2);
-short2 __ovld __cnfn convert_short2_rtz(long2);
-short2 __ovld __cnfn convert_short2_sat_rtz(long2);
-short2 __ovld __cnfn convert_short2_rtp(long2);
-short2 __ovld __cnfn convert_short2_sat_rtp(long2);
-short2 __ovld __cnfn convert_short2_rtn(long2);
-short2 __ovld __cnfn convert_short2_sat_rtn(long2);
-short2 __ovld __cnfn convert_short2(long2);
-short2 __ovld __cnfn convert_short2_sat(long2);
-short2 __ovld __cnfn convert_short2_rte(ulong2);
-short2 __ovld __cnfn convert_short2_sat_rte(ulong2);
-short2 __ovld __cnfn convert_short2_rtz(ulong2);
-short2 __ovld __cnfn convert_short2_sat_rtz(ulong2);
-short2 __ovld __cnfn convert_short2_rtp(ulong2);
-short2 __ovld __cnfn convert_short2_sat_rtp(ulong2);
-short2 __ovld __cnfn convert_short2_rtn(ulong2);
-short2 __ovld __cnfn convert_short2_sat_rtn(ulong2);
-short2 __ovld __cnfn convert_short2(ulong2);
-short2 __ovld __cnfn convert_short2_sat(ulong2);
-short2 __ovld __cnfn convert_short2_rte(float2);
-short2 __ovld __cnfn convert_short2_sat_rte(float2);
-short2 __ovld __cnfn convert_short2_rtz(float2);
-short2 __ovld __cnfn convert_short2_sat_rtz(float2);
-short2 __ovld __cnfn convert_short2_rtp(float2);
-short2 __ovld __cnfn convert_short2_sat_rtp(float2);
-short2 __ovld __cnfn convert_short2_rtn(float2);
-short2 __ovld __cnfn convert_short2_sat_rtn(float2);
-short2 __ovld __cnfn convert_short2(float2);
-short2 __ovld __cnfn convert_short2_sat(float2);
-ushort2 __ovld __cnfn convert_ushort2_rte(char2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(char2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(char2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(char2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(char2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(char2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(char2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(char2);
-ushort2 __ovld __cnfn convert_ushort2(char2);
-ushort2 __ovld __cnfn convert_ushort2_sat(char2);
-ushort2 __ovld __cnfn convert_ushort2_rte(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(uchar2);
-ushort2 __ovld __cnfn convert_ushort2(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_sat(uchar2);
-ushort2 __ovld __cnfn convert_ushort2_rte(short2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(short2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(short2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(short2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(short2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(short2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(short2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(short2);
-ushort2 __ovld __cnfn convert_ushort2(short2);
-ushort2 __ovld __cnfn convert_ushort2_sat(short2);
-ushort2 __ovld __cnfn convert_ushort2_rte(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(ushort2);
-ushort2 __ovld __cnfn convert_ushort2(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_sat(ushort2);
-ushort2 __ovld __cnfn convert_ushort2_rte(int2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(int2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(int2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(int2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(int2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(int2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(int2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(int2);
-ushort2 __ovld __cnfn convert_ushort2(int2);
-ushort2 __ovld __cnfn convert_ushort2_sat(int2);
-ushort2 __ovld __cnfn convert_ushort2_rte(uint2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(uint2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(uint2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(uint2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(uint2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(uint2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(uint2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(uint2);
-ushort2 __ovld __cnfn convert_ushort2(uint2);
-ushort2 __ovld __cnfn convert_ushort2_sat(uint2);
-ushort2 __ovld __cnfn convert_ushort2_rte(long2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(long2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(long2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(long2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(long2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(long2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(long2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(long2);
-ushort2 __ovld __cnfn convert_ushort2(long2);
-ushort2 __ovld __cnfn convert_ushort2_sat(long2);
-ushort2 __ovld __cnfn convert_ushort2_rte(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(ulong2);
-ushort2 __ovld __cnfn convert_ushort2(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_sat(ulong2);
-ushort2 __ovld __cnfn convert_ushort2_rte(float2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(float2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(float2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(float2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(float2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(float2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(float2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(float2);
-ushort2 __ovld __cnfn convert_ushort2(float2);
-ushort2 __ovld __cnfn convert_ushort2_sat(float2);
-int2 __ovld __cnfn convert_int2_rte(char2);
-int2 __ovld __cnfn convert_int2_sat_rte(char2);
-int2 __ovld __cnfn convert_int2_rtz(char2);
-int2 __ovld __cnfn convert_int2_sat_rtz(char2);
-int2 __ovld __cnfn convert_int2_rtp(char2);
-int2 __ovld __cnfn convert_int2_sat_rtp(char2);
-int2 __ovld __cnfn convert_int2_rtn(char2);
-int2 __ovld __cnfn convert_int2_sat_rtn(char2);
-int2 __ovld __cnfn convert_int2(char2);
-int2 __ovld __cnfn convert_int2_sat(char2);
-int2 __ovld __cnfn convert_int2_rte(uchar2);
-int2 __ovld __cnfn convert_int2_sat_rte(uchar2);
-int2 __ovld __cnfn convert_int2_rtz(uchar2);
-int2 __ovld __cnfn convert_int2_sat_rtz(uchar2);
-int2 __ovld __cnfn convert_int2_rtp(uchar2);
-int2 __ovld __cnfn convert_int2_sat_rtp(uchar2);
-int2 __ovld __cnfn convert_int2_rtn(uchar2);
-int2 __ovld __cnfn convert_int2_sat_rtn(uchar2);
-int2 __ovld __cnfn convert_int2(uchar2);
-int2 __ovld __cnfn convert_int2_sat(uchar2);
-int2 __ovld __cnfn convert_int2_rte(short2);
-int2 __ovld __cnfn convert_int2_sat_rte(short2);
-int2 __ovld __cnfn convert_int2_rtz(short2);
-int2 __ovld __cnfn convert_int2_sat_rtz(short2);
-int2 __ovld __cnfn convert_int2_rtp(short2);
-int2 __ovld __cnfn convert_int2_sat_rtp(short2);
-int2 __ovld __cnfn convert_int2_rtn(short2);
-int2 __ovld __cnfn convert_int2_sat_rtn(short2);
-int2 __ovld __cnfn convert_int2(short2);
-int2 __ovld __cnfn convert_int2_sat(short2);
-int2 __ovld __cnfn convert_int2_rte(ushort2);
-int2 __ovld __cnfn convert_int2_sat_rte(ushort2);
-int2 __ovld __cnfn convert_int2_rtz(ushort2);
-int2 __ovld __cnfn convert_int2_sat_rtz(ushort2);
-int2 __ovld __cnfn convert_int2_rtp(ushort2);
-int2 __ovld __cnfn convert_int2_sat_rtp(ushort2);
-int2 __ovld __cnfn convert_int2_rtn(ushort2);
-int2 __ovld __cnfn convert_int2_sat_rtn(ushort2);
-int2 __ovld __cnfn convert_int2(ushort2);
-int2 __ovld __cnfn convert_int2_sat(ushort2);
-int2 __ovld __cnfn convert_int2_rte(int2);
-int2 __ovld __cnfn convert_int2_sat_rte(int2);
-int2 __ovld __cnfn convert_int2_rtz(int2);
-int2 __ovld __cnfn convert_int2_sat_rtz(int2);
-int2 __ovld __cnfn convert_int2_rtp(int2);
-int2 __ovld __cnfn convert_int2_sat_rtp(int2);
-int2 __ovld __cnfn convert_int2_rtn(int2);
-int2 __ovld __cnfn convert_int2_sat_rtn(int2);
-int2 __ovld __cnfn convert_int2(int2);
-int2 __ovld __cnfn convert_int2_sat(int2);
-int2 __ovld __cnfn convert_int2_rte(uint2);
-int2 __ovld __cnfn convert_int2_sat_rte(uint2);
-int2 __ovld __cnfn convert_int2_rtz(uint2);
-int2 __ovld __cnfn convert_int2_sat_rtz(uint2);
-int2 __ovld __cnfn convert_int2_rtp(uint2);
-int2 __ovld __cnfn convert_int2_sat_rtp(uint2);
-int2 __ovld __cnfn convert_int2_rtn(uint2);
-int2 __ovld __cnfn convert_int2_sat_rtn(uint2);
-int2 __ovld __cnfn convert_int2(uint2);
-int2 __ovld __cnfn convert_int2_sat(uint2);
-int2 __ovld __cnfn convert_int2_rte(long2);
-int2 __ovld __cnfn convert_int2_sat_rte(long2);
-int2 __ovld __cnfn convert_int2_rtz(long2);
-int2 __ovld __cnfn convert_int2_sat_rtz(long2);
-int2 __ovld __cnfn convert_int2_rtp(long2);
-int2 __ovld __cnfn convert_int2_sat_rtp(long2);
-int2 __ovld __cnfn convert_int2_rtn(long2);
-int2 __ovld __cnfn convert_int2_sat_rtn(long2);
-int2 __ovld __cnfn convert_int2(long2);
-int2 __ovld __cnfn convert_int2_sat(long2);
-int2 __ovld __cnfn convert_int2_rte(ulong2);
-int2 __ovld __cnfn convert_int2_sat_rte(ulong2);
-int2 __ovld __cnfn convert_int2_rtz(ulong2);
-int2 __ovld __cnfn convert_int2_sat_rtz(ulong2);
-int2 __ovld __cnfn convert_int2_rtp(ulong2);
-int2 __ovld __cnfn convert_int2_sat_rtp(ulong2);
-int2 __ovld __cnfn convert_int2_rtn(ulong2);
-int2 __ovld __cnfn convert_int2_sat_rtn(ulong2);
-int2 __ovld __cnfn convert_int2(ulong2);
-int2 __ovld __cnfn convert_int2_sat(ulong2);
-int2 __ovld __cnfn convert_int2_rte(float2);
-int2 __ovld __cnfn convert_int2_sat_rte(float2);
-int2 __ovld __cnfn convert_int2_rtz(float2);
-int2 __ovld __cnfn convert_int2_sat_rtz(float2);
-int2 __ovld __cnfn convert_int2_rtp(float2);
-int2 __ovld __cnfn convert_int2_sat_rtp(float2);
-int2 __ovld __cnfn convert_int2_rtn(float2);
-int2 __ovld __cnfn convert_int2_sat_rtn(float2);
-int2 __ovld __cnfn convert_int2(float2);
-int2 __ovld __cnfn convert_int2_sat(float2);
-uint2 __ovld __cnfn convert_uint2_rte(char2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(char2);
-uint2 __ovld __cnfn convert_uint2_rtz(char2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(char2);
-uint2 __ovld __cnfn convert_uint2_rtp(char2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(char2);
-uint2 __ovld __cnfn convert_uint2_rtn(char2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(char2);
-uint2 __ovld __cnfn convert_uint2(char2);
-uint2 __ovld __cnfn convert_uint2_sat(char2);
-uint2 __ovld __cnfn convert_uint2_rte(uchar2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(uchar2);
-uint2 __ovld __cnfn convert_uint2_rtz(uchar2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(uchar2);
-uint2 __ovld __cnfn convert_uint2_rtp(uchar2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(uchar2);
-uint2 __ovld __cnfn convert_uint2_rtn(uchar2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(uchar2);
-uint2 __ovld __cnfn convert_uint2(uchar2);
-uint2 __ovld __cnfn convert_uint2_sat(uchar2);
-uint2 __ovld __cnfn convert_uint2_rte(short2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(short2);
-uint2 __ovld __cnfn convert_uint2_rtz(short2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(short2);
-uint2 __ovld __cnfn convert_uint2_rtp(short2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(short2);
-uint2 __ovld __cnfn convert_uint2_rtn(short2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(short2);
-uint2 __ovld __cnfn convert_uint2(short2);
-uint2 __ovld __cnfn convert_uint2_sat(short2);
-uint2 __ovld __cnfn convert_uint2_rte(ushort2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(ushort2);
-uint2 __ovld __cnfn convert_uint2_rtz(ushort2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(ushort2);
-uint2 __ovld __cnfn convert_uint2_rtp(ushort2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(ushort2);
-uint2 __ovld __cnfn convert_uint2_rtn(ushort2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(ushort2);
-uint2 __ovld __cnfn convert_uint2(ushort2);
-uint2 __ovld __cnfn convert_uint2_sat(ushort2);
-uint2 __ovld __cnfn convert_uint2_rte(int2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(int2);
-uint2 __ovld __cnfn convert_uint2_rtz(int2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(int2);
-uint2 __ovld __cnfn convert_uint2_rtp(int2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(int2);
-uint2 __ovld __cnfn convert_uint2_rtn(int2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(int2);
-uint2 __ovld __cnfn convert_uint2(int2);
-uint2 __ovld __cnfn convert_uint2_sat(int2);
-uint2 __ovld __cnfn convert_uint2_rte(uint2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(uint2);
-uint2 __ovld __cnfn convert_uint2_rtz(uint2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(uint2);
-uint2 __ovld __cnfn convert_uint2_rtp(uint2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(uint2);
-uint2 __ovld __cnfn convert_uint2_rtn(uint2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(uint2);
-uint2 __ovld __cnfn convert_uint2(uint2);
-uint2 __ovld __cnfn convert_uint2_sat(uint2);
-uint2 __ovld __cnfn convert_uint2_rte(long2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(long2);
-uint2 __ovld __cnfn convert_uint2_rtz(long2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(long2);
-uint2 __ovld __cnfn convert_uint2_rtp(long2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(long2);
-uint2 __ovld __cnfn convert_uint2_rtn(long2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(long2);
-uint2 __ovld __cnfn convert_uint2(long2);
-uint2 __ovld __cnfn convert_uint2_sat(long2);
-uint2 __ovld __cnfn convert_uint2_rte(ulong2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(ulong2);
-uint2 __ovld __cnfn convert_uint2_rtz(ulong2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(ulong2);
-uint2 __ovld __cnfn convert_uint2_rtp(ulong2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(ulong2);
-uint2 __ovld __cnfn convert_uint2_rtn(ulong2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(ulong2);
-uint2 __ovld __cnfn convert_uint2(ulong2);
-uint2 __ovld __cnfn convert_uint2_sat(ulong2);
-uint2 __ovld __cnfn convert_uint2_rte(float2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(float2);
-uint2 __ovld __cnfn convert_uint2_rtz(float2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(float2);
-uint2 __ovld __cnfn convert_uint2_rtp(float2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(float2);
-uint2 __ovld __cnfn convert_uint2_rtn(float2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(float2);
-uint2 __ovld __cnfn convert_uint2(float2);
-uint2 __ovld __cnfn convert_uint2_sat(float2);
-long2 __ovld __cnfn convert_long2_rte(char2);
-long2 __ovld __cnfn convert_long2_sat_rte(char2);
-long2 __ovld __cnfn convert_long2_rtz(char2);
-long2 __ovld __cnfn convert_long2_sat_rtz(char2);
-long2 __ovld __cnfn convert_long2_rtp(char2);
-long2 __ovld __cnfn convert_long2_sat_rtp(char2);
-long2 __ovld __cnfn convert_long2_rtn(char2);
-long2 __ovld __cnfn convert_long2_sat_rtn(char2);
-long2 __ovld __cnfn convert_long2(char2);
-long2 __ovld __cnfn convert_long2_sat(char2);
-long2 __ovld __cnfn convert_long2_rte(uchar2);
-long2 __ovld __cnfn convert_long2_sat_rte(uchar2);
-long2 __ovld __cnfn convert_long2_rtz(uchar2);
-long2 __ovld __cnfn convert_long2_sat_rtz(uchar2);
-long2 __ovld __cnfn convert_long2_rtp(uchar2);
-long2 __ovld __cnfn convert_long2_sat_rtp(uchar2);
-long2 __ovld __cnfn convert_long2_rtn(uchar2);
-long2 __ovld __cnfn convert_long2_sat_rtn(uchar2);
-long2 __ovld __cnfn convert_long2(uchar2);
-long2 __ovld __cnfn convert_long2_sat(uchar2);
-long2 __ovld __cnfn convert_long2_rte(short2);
-long2 __ovld __cnfn convert_long2_sat_rte(short2);
-long2 __ovld __cnfn convert_long2_rtz(short2);
-long2 __ovld __cnfn convert_long2_sat_rtz(short2);
-long2 __ovld __cnfn convert_long2_rtp(short2);
-long2 __ovld __cnfn convert_long2_sat_rtp(short2);
-long2 __ovld __cnfn convert_long2_rtn(short2);
-long2 __ovld __cnfn convert_long2_sat_rtn(short2);
-long2 __ovld __cnfn convert_long2(short2);
-long2 __ovld __cnfn convert_long2_sat(short2);
-long2 __ovld __cnfn convert_long2_rte(ushort2);
-long2 __ovld __cnfn convert_long2_sat_rte(ushort2);
-long2 __ovld __cnfn convert_long2_rtz(ushort2);
-long2 __ovld __cnfn convert_long2_sat_rtz(ushort2);
-long2 __ovld __cnfn convert_long2_rtp(ushort2);
-long2 __ovld __cnfn convert_long2_sat_rtp(ushort2);
-long2 __ovld __cnfn convert_long2_rtn(ushort2);
-long2 __ovld __cnfn convert_long2_sat_rtn(ushort2);
-long2 __ovld __cnfn convert_long2(ushort2);
-long2 __ovld __cnfn convert_long2_sat(ushort2);
-long2 __ovld __cnfn convert_long2_rte(int2);
-long2 __ovld __cnfn convert_long2_sat_rte(int2);
-long2 __ovld __cnfn convert_long2_rtz(int2);
-long2 __ovld __cnfn convert_long2_sat_rtz(int2);
-long2 __ovld __cnfn convert_long2_rtp(int2);
-long2 __ovld __cnfn convert_long2_sat_rtp(int2);
-long2 __ovld __cnfn convert_long2_rtn(int2);
-long2 __ovld __cnfn convert_long2_sat_rtn(int2);
-long2 __ovld __cnfn convert_long2(int2);
-long2 __ovld __cnfn convert_long2_sat(int2);
-long2 __ovld __cnfn convert_long2_rte(uint2);
-long2 __ovld __cnfn convert_long2_sat_rte(uint2);
-long2 __ovld __cnfn convert_long2_rtz(uint2);
-long2 __ovld __cnfn convert_long2_sat_rtz(uint2);
-long2 __ovld __cnfn convert_long2_rtp(uint2);
-long2 __ovld __cnfn convert_long2_sat_rtp(uint2);
-long2 __ovld __cnfn convert_long2_rtn(uint2);
-long2 __ovld __cnfn convert_long2_sat_rtn(uint2);
-long2 __ovld __cnfn convert_long2(uint2);
-long2 __ovld __cnfn convert_long2_sat(uint2);
-long2 __ovld __cnfn convert_long2_rte(long2);
-long2 __ovld __cnfn convert_long2_sat_rte(long2);
-long2 __ovld __cnfn convert_long2_rtz(long2);
-long2 __ovld __cnfn convert_long2_sat_rtz(long2);
-long2 __ovld __cnfn convert_long2_rtp(long2);
-long2 __ovld __cnfn convert_long2_sat_rtp(long2);
-long2 __ovld __cnfn convert_long2_rtn(long2);
-long2 __ovld __cnfn convert_long2_sat_rtn(long2);
-long2 __ovld __cnfn convert_long2(long2);
-long2 __ovld __cnfn convert_long2_sat(long2);
-long2 __ovld __cnfn convert_long2_rte(ulong2);
-long2 __ovld __cnfn convert_long2_sat_rte(ulong2);
-long2 __ovld __cnfn convert_long2_rtz(ulong2);
-long2 __ovld __cnfn convert_long2_sat_rtz(ulong2);
-long2 __ovld __cnfn convert_long2_rtp(ulong2);
-long2 __ovld __cnfn convert_long2_sat_rtp(ulong2);
-long2 __ovld __cnfn convert_long2_rtn(ulong2);
-long2 __ovld __cnfn convert_long2_sat_rtn(ulong2);
-long2 __ovld __cnfn convert_long2(ulong2);
-long2 __ovld __cnfn convert_long2_sat(ulong2);
-long2 __ovld __cnfn convert_long2_rte(float2);
-long2 __ovld __cnfn convert_long2_sat_rte(float2);
-long2 __ovld __cnfn convert_long2_rtz(float2);
-long2 __ovld __cnfn convert_long2_sat_rtz(float2);
-long2 __ovld __cnfn convert_long2_rtp(float2);
-long2 __ovld __cnfn convert_long2_sat_rtp(float2);
-long2 __ovld __cnfn convert_long2_rtn(float2);
-long2 __ovld __cnfn convert_long2_sat_rtn(float2);
-long2 __ovld __cnfn convert_long2(float2);
-long2 __ovld __cnfn convert_long2_sat(float2);
-ulong2 __ovld __cnfn convert_ulong2_rte(char2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(char2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(char2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(char2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(char2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(char2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(char2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(char2);
-ulong2 __ovld __cnfn convert_ulong2(char2);
-ulong2 __ovld __cnfn convert_ulong2_sat(char2);
-ulong2 __ovld __cnfn convert_ulong2_rte(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(uchar2);
-ulong2 __ovld __cnfn convert_ulong2(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_sat(uchar2);
-ulong2 __ovld __cnfn convert_ulong2_rte(short2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(short2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(short2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(short2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(short2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(short2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(short2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(short2);
-ulong2 __ovld __cnfn convert_ulong2(short2);
-ulong2 __ovld __cnfn convert_ulong2_sat(short2);
-ulong2 __ovld __cnfn convert_ulong2_rte(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(ushort2);
-ulong2 __ovld __cnfn convert_ulong2(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_sat(ushort2);
-ulong2 __ovld __cnfn convert_ulong2_rte(int2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(int2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(int2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(int2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(int2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(int2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(int2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(int2);
-ulong2 __ovld __cnfn convert_ulong2(int2);
-ulong2 __ovld __cnfn convert_ulong2_sat(int2);
-ulong2 __ovld __cnfn convert_ulong2_rte(uint2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(uint2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(uint2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(uint2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(uint2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(uint2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(uint2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(uint2);
-ulong2 __ovld __cnfn convert_ulong2(uint2);
-ulong2 __ovld __cnfn convert_ulong2_sat(uint2);
-ulong2 __ovld __cnfn convert_ulong2_rte(long2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(long2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(long2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(long2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(long2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(long2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(long2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(long2);
-ulong2 __ovld __cnfn convert_ulong2(long2);
-ulong2 __ovld __cnfn convert_ulong2_sat(long2);
-ulong2 __ovld __cnfn convert_ulong2_rte(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(ulong2);
-ulong2 __ovld __cnfn convert_ulong2(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_sat(ulong2);
-ulong2 __ovld __cnfn convert_ulong2_rte(float2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(float2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(float2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(float2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(float2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(float2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(float2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(float2);
-ulong2 __ovld __cnfn convert_ulong2(float2);
-ulong2 __ovld __cnfn convert_ulong2_sat(float2);
-float2 __ovld __cnfn convert_float2_rte(char2);
-float2 __ovld __cnfn convert_float2_rtz(char2);
-float2 __ovld __cnfn convert_float2_rtp(char2);
-float2 __ovld __cnfn convert_float2_rtn(char2);
-float2 __ovld __cnfn convert_float2(char2);
-float2 __ovld __cnfn convert_float2_rte(uchar2);
-float2 __ovld __cnfn convert_float2_rtz(uchar2);
-float2 __ovld __cnfn convert_float2_rtp(uchar2);
-float2 __ovld __cnfn convert_float2_rtn(uchar2);
-float2 __ovld __cnfn convert_float2(uchar2);
-float2 __ovld __cnfn convert_float2_rte(short2);
-float2 __ovld __cnfn convert_float2_rtz(short2);
-float2 __ovld __cnfn convert_float2_rtp(short2);
-float2 __ovld __cnfn convert_float2_rtn(short2);
-float2 __ovld __cnfn convert_float2(short2);
-float2 __ovld __cnfn convert_float2_rte(ushort2);
-float2 __ovld __cnfn convert_float2_rtz(ushort2);
-float2 __ovld __cnfn convert_float2_rtp(ushort2);
-float2 __ovld __cnfn convert_float2_rtn(ushort2);
-float2 __ovld __cnfn convert_float2(ushort2);
-float2 __ovld __cnfn convert_float2_rte(int2);
-float2 __ovld __cnfn convert_float2_rtz(int2);
-float2 __ovld __cnfn convert_float2_rtp(int2);
-float2 __ovld __cnfn convert_float2_rtn(int2);
-float2 __ovld __cnfn convert_float2(int2);
-float2 __ovld __cnfn convert_float2_rte(uint2);
-float2 __ovld __cnfn convert_float2_rtz(uint2);
-float2 __ovld __cnfn convert_float2_rtp(uint2);
-float2 __ovld __cnfn convert_float2_rtn(uint2);
-float2 __ovld __cnfn convert_float2(uint2);
-float2 __ovld __cnfn convert_float2_rte(long2);
-float2 __ovld __cnfn convert_float2_rtz(long2);
-float2 __ovld __cnfn convert_float2_rtp(long2);
-float2 __ovld __cnfn convert_float2_rtn(long2);
-float2 __ovld __cnfn convert_float2(long2);
-float2 __ovld __cnfn convert_float2_rte(ulong2);
-float2 __ovld __cnfn convert_float2_rtz(ulong2);
-float2 __ovld __cnfn convert_float2_rtp(ulong2);
-float2 __ovld __cnfn convert_float2_rtn(ulong2);
-float2 __ovld __cnfn convert_float2(ulong2);
-float2 __ovld __cnfn convert_float2_rte(float2);
-float2 __ovld __cnfn convert_float2_rtz(float2);
-float2 __ovld __cnfn convert_float2_rtp(float2);
-float2 __ovld __cnfn convert_float2_rtn(float2);
-float2 __ovld __cnfn convert_float2(float2);
-char3 __ovld __cnfn convert_char3_rte(char3);
-char3 __ovld __cnfn convert_char3_sat_rte(char3);
-char3 __ovld __cnfn convert_char3_rtz(char3);
-char3 __ovld __cnfn convert_char3_sat_rtz(char3);
-char3 __ovld __cnfn convert_char3_rtp(char3);
-char3 __ovld __cnfn convert_char3_sat_rtp(char3);
-char3 __ovld __cnfn convert_char3_rtn(char3);
-char3 __ovld __cnfn convert_char3_sat_rtn(char3);
-char3 __ovld __cnfn convert_char3(char3);
-char3 __ovld __cnfn convert_char3_sat(char3);
-char3 __ovld __cnfn convert_char3_rte(uchar3);
-char3 __ovld __cnfn convert_char3_sat_rte(uchar3);
-char3 __ovld __cnfn convert_char3_rtz(uchar3);
-char3 __ovld __cnfn convert_char3_sat_rtz(uchar3);
-char3 __ovld __cnfn convert_char3_rtp(uchar3);
-char3 __ovld __cnfn convert_char3_sat_rtp(uchar3);
-char3 __ovld __cnfn convert_char3_rtn(uchar3);
-char3 __ovld __cnfn convert_char3_sat_rtn(uchar3);
-char3 __ovld __cnfn convert_char3(uchar3);
-char3 __ovld __cnfn convert_char3_sat(uchar3);
-char3 __ovld __cnfn convert_char3_rte(short3);
-char3 __ovld __cnfn convert_char3_sat_rte(short3);
-char3 __ovld __cnfn convert_char3_rtz(short3);
-char3 __ovld __cnfn convert_char3_sat_rtz(short3);
-char3 __ovld __cnfn convert_char3_rtp(short3);
-char3 __ovld __cnfn convert_char3_sat_rtp(short3);
-char3 __ovld __cnfn convert_char3_rtn(short3);
-char3 __ovld __cnfn convert_char3_sat_rtn(short3);
-char3 __ovld __cnfn convert_char3(short3);
-char3 __ovld __cnfn convert_char3_sat(short3);
-char3 __ovld __cnfn convert_char3_rte(ushort3);
-char3 __ovld __cnfn convert_char3_sat_rte(ushort3);
-char3 __ovld __cnfn convert_char3_rtz(ushort3);
-char3 __ovld __cnfn convert_char3_sat_rtz(ushort3);
-char3 __ovld __cnfn convert_char3_rtp(ushort3);
-char3 __ovld __cnfn convert_char3_sat_rtp(ushort3);
-char3 __ovld __cnfn convert_char3_rtn(ushort3);
-char3 __ovld __cnfn convert_char3_sat_rtn(ushort3);
-char3 __ovld __cnfn convert_char3(ushort3);
-char3 __ovld __cnfn convert_char3_sat(ushort3);
-char3 __ovld __cnfn convert_char3_rte(int3);
-char3 __ovld __cnfn convert_char3_sat_rte(int3);
-char3 __ovld __cnfn convert_char3_rtz(int3);
-char3 __ovld __cnfn convert_char3_sat_rtz(int3);
-char3 __ovld __cnfn convert_char3_rtp(int3);
-char3 __ovld __cnfn convert_char3_sat_rtp(int3);
-char3 __ovld __cnfn convert_char3_rtn(int3);
-char3 __ovld __cnfn convert_char3_sat_rtn(int3);
-char3 __ovld __cnfn convert_char3(int3);
-char3 __ovld __cnfn convert_char3_sat(int3);
-char3 __ovld __cnfn convert_char3_rte(uint3);
-char3 __ovld __cnfn convert_char3_sat_rte(uint3);
-char3 __ovld __cnfn convert_char3_rtz(uint3);
-char3 __ovld __cnfn convert_char3_sat_rtz(uint3);
-char3 __ovld __cnfn convert_char3_rtp(uint3);
-char3 __ovld __cnfn convert_char3_sat_rtp(uint3);
-char3 __ovld __cnfn convert_char3_rtn(uint3);
-char3 __ovld __cnfn convert_char3_sat_rtn(uint3);
-char3 __ovld __cnfn convert_char3(uint3);
-char3 __ovld __cnfn convert_char3_sat(uint3);
-char3 __ovld __cnfn convert_char3_rte(long3);
-char3 __ovld __cnfn convert_char3_sat_rte(long3);
-char3 __ovld __cnfn convert_char3_rtz(long3);
-char3 __ovld __cnfn convert_char3_sat_rtz(long3);
-char3 __ovld __cnfn convert_char3_rtp(long3);
-char3 __ovld __cnfn convert_char3_sat_rtp(long3);
-char3 __ovld __cnfn convert_char3_rtn(long3);
-char3 __ovld __cnfn convert_char3_sat_rtn(long3);
-char3 __ovld __cnfn convert_char3(long3);
-char3 __ovld __cnfn convert_char3_sat(long3);
-char3 __ovld __cnfn convert_char3_rte(ulong3);
-char3 __ovld __cnfn convert_char3_sat_rte(ulong3);
-char3 __ovld __cnfn convert_char3_rtz(ulong3);
-char3 __ovld __cnfn convert_char3_sat_rtz(ulong3);
-char3 __ovld __cnfn convert_char3_rtp(ulong3);
-char3 __ovld __cnfn convert_char3_sat_rtp(ulong3);
-char3 __ovld __cnfn convert_char3_rtn(ulong3);
-char3 __ovld __cnfn convert_char3_sat_rtn(ulong3);
-char3 __ovld __cnfn convert_char3(ulong3);
-char3 __ovld __cnfn convert_char3_sat(ulong3);
-char3 __ovld __cnfn convert_char3_rte(float3);
-char3 __ovld __cnfn convert_char3_sat_rte(float3);
-char3 __ovld __cnfn convert_char3_rtz(float3);
-char3 __ovld __cnfn convert_char3_sat_rtz(float3);
-char3 __ovld __cnfn convert_char3_rtp(float3);
-char3 __ovld __cnfn convert_char3_sat_rtp(float3);
-char3 __ovld __cnfn convert_char3_rtn(float3);
-char3 __ovld __cnfn convert_char3_sat_rtn(float3);
-char3 __ovld __cnfn convert_char3(float3);
-char3 __ovld __cnfn convert_char3_sat(float3);
-uchar3 __ovld __cnfn convert_uchar3_rte(char3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(char3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(char3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(char3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(char3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(char3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(char3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(char3);
-uchar3 __ovld __cnfn convert_uchar3(char3);
-uchar3 __ovld __cnfn convert_uchar3_sat(char3);
-uchar3 __ovld __cnfn convert_uchar3_rte(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(uchar3);
-uchar3 __ovld __cnfn convert_uchar3(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_sat(uchar3);
-uchar3 __ovld __cnfn convert_uchar3_rte(short3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(short3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(short3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(short3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(short3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(short3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(short3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(short3);
-uchar3 __ovld __cnfn convert_uchar3(short3);
-uchar3 __ovld __cnfn convert_uchar3_sat(short3);
-uchar3 __ovld __cnfn convert_uchar3_rte(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(ushort3);
-uchar3 __ovld __cnfn convert_uchar3(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_sat(ushort3);
-uchar3 __ovld __cnfn convert_uchar3_rte(int3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(int3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(int3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(int3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(int3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(int3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(int3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(int3);
-uchar3 __ovld __cnfn convert_uchar3(int3);
-uchar3 __ovld __cnfn convert_uchar3_sat(int3);
-uchar3 __ovld __cnfn convert_uchar3_rte(uint3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(uint3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(uint3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(uint3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(uint3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(uint3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(uint3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(uint3);
-uchar3 __ovld __cnfn convert_uchar3(uint3);
-uchar3 __ovld __cnfn convert_uchar3_sat(uint3);
-uchar3 __ovld __cnfn convert_uchar3_rte(long3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(long3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(long3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(long3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(long3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(long3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(long3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(long3);
-uchar3 __ovld __cnfn convert_uchar3(long3);
-uchar3 __ovld __cnfn convert_uchar3_sat(long3);
-uchar3 __ovld __cnfn convert_uchar3_rte(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(ulong3);
-uchar3 __ovld __cnfn convert_uchar3(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_sat(ulong3);
-uchar3 __ovld __cnfn convert_uchar3_rte(float3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(float3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(float3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(float3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(float3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(float3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(float3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(float3);
-uchar3 __ovld __cnfn convert_uchar3(float3);
-uchar3 __ovld __cnfn convert_uchar3_sat(float3);
-short3 __ovld __cnfn convert_short3_rte(char3);
-short3 __ovld __cnfn convert_short3_sat_rte(char3);
-short3 __ovld __cnfn convert_short3_rtz(char3);
-short3 __ovld __cnfn convert_short3_sat_rtz(char3);
-short3 __ovld __cnfn convert_short3_rtp(char3);
-short3 __ovld __cnfn convert_short3_sat_rtp(char3);
-short3 __ovld __cnfn convert_short3_rtn(char3);
-short3 __ovld __cnfn convert_short3_sat_rtn(char3);
-short3 __ovld __cnfn convert_short3(char3);
-short3 __ovld __cnfn convert_short3_sat(char3);
-short3 __ovld __cnfn convert_short3_rte(uchar3);
-short3 __ovld __cnfn convert_short3_sat_rte(uchar3);
-short3 __ovld __cnfn convert_short3_rtz(uchar3);
-short3 __ovld __cnfn convert_short3_sat_rtz(uchar3);
-short3 __ovld __cnfn convert_short3_rtp(uchar3);
-short3 __ovld __cnfn convert_short3_sat_rtp(uchar3);
-short3 __ovld __cnfn convert_short3_rtn(uchar3);
-short3 __ovld __cnfn convert_short3_sat_rtn(uchar3);
-short3 __ovld __cnfn convert_short3(uchar3);
-short3 __ovld __cnfn convert_short3_sat(uchar3);
-short3 __ovld __cnfn convert_short3_rte(short3);
-short3 __ovld __cnfn convert_short3_sat_rte(short3);
-short3 __ovld __cnfn convert_short3_rtz(short3);
-short3 __ovld __cnfn convert_short3_sat_rtz(short3);
-short3 __ovld __cnfn convert_short3_rtp(short3);
-short3 __ovld __cnfn convert_short3_sat_rtp(short3);
-short3 __ovld __cnfn convert_short3_rtn(short3);
-short3 __ovld __cnfn convert_short3_sat_rtn(short3);
-short3 __ovld __cnfn convert_short3(short3);
-short3 __ovld __cnfn convert_short3_sat(short3);
-short3 __ovld __cnfn convert_short3_rte(ushort3);
-short3 __ovld __cnfn convert_short3_sat_rte(ushort3);
-short3 __ovld __cnfn convert_short3_rtz(ushort3);
-short3 __ovld __cnfn convert_short3_sat_rtz(ushort3);
-short3 __ovld __cnfn convert_short3_rtp(ushort3);
-short3 __ovld __cnfn convert_short3_sat_rtp(ushort3);
-short3 __ovld __cnfn convert_short3_rtn(ushort3);
-short3 __ovld __cnfn convert_short3_sat_rtn(ushort3);
-short3 __ovld __cnfn convert_short3(ushort3);
-short3 __ovld __cnfn convert_short3_sat(ushort3);
-short3 __ovld __cnfn convert_short3_rte(int3);
-short3 __ovld __cnfn convert_short3_sat_rte(int3);
-short3 __ovld __cnfn convert_short3_rtz(int3);
-short3 __ovld __cnfn convert_short3_sat_rtz(int3);
-short3 __ovld __cnfn convert_short3_rtp(int3);
-short3 __ovld __cnfn convert_short3_sat_rtp(int3);
-short3 __ovld __cnfn convert_short3_rtn(int3);
-short3 __ovld __cnfn convert_short3_sat_rtn(int3);
-short3 __ovld __cnfn convert_short3(int3);
-short3 __ovld __cnfn convert_short3_sat(int3);
-short3 __ovld __cnfn convert_short3_rte(uint3);
-short3 __ovld __cnfn convert_short3_sat_rte(uint3);
-short3 __ovld __cnfn convert_short3_rtz(uint3);
-short3 __ovld __cnfn convert_short3_sat_rtz(uint3);
-short3 __ovld __cnfn convert_short3_rtp(uint3);
-short3 __ovld __cnfn convert_short3_sat_rtp(uint3);
-short3 __ovld __cnfn convert_short3_rtn(uint3);
-short3 __ovld __cnfn convert_short3_sat_rtn(uint3);
-short3 __ovld __cnfn convert_short3(uint3);
-short3 __ovld __cnfn convert_short3_sat(uint3);
-short3 __ovld __cnfn convert_short3_rte(long3);
-short3 __ovld __cnfn convert_short3_sat_rte(long3);
-short3 __ovld __cnfn convert_short3_rtz(long3);
-short3 __ovld __cnfn convert_short3_sat_rtz(long3);
-short3 __ovld __cnfn convert_short3_rtp(long3);
-short3 __ovld __cnfn convert_short3_sat_rtp(long3);
-short3 __ovld __cnfn convert_short3_rtn(long3);
-short3 __ovld __cnfn convert_short3_sat_rtn(long3);
-short3 __ovld __cnfn convert_short3(long3);
-short3 __ovld __cnfn convert_short3_sat(long3);
-short3 __ovld __cnfn convert_short3_rte(ulong3);
-short3 __ovld __cnfn convert_short3_sat_rte(ulong3);
-short3 __ovld __cnfn convert_short3_rtz(ulong3);
-short3 __ovld __cnfn convert_short3_sat_rtz(ulong3);
-short3 __ovld __cnfn convert_short3_rtp(ulong3);
-short3 __ovld __cnfn convert_short3_sat_rtp(ulong3);
-short3 __ovld __cnfn convert_short3_rtn(ulong3);
-short3 __ovld __cnfn convert_short3_sat_rtn(ulong3);
-short3 __ovld __cnfn convert_short3(ulong3);
-short3 __ovld __cnfn convert_short3_sat(ulong3);
-short3 __ovld __cnfn convert_short3_rte(float3);
-short3 __ovld __cnfn convert_short3_sat_rte(float3);
-short3 __ovld __cnfn convert_short3_rtz(float3);
-short3 __ovld __cnfn convert_short3_sat_rtz(float3);
-short3 __ovld __cnfn convert_short3_rtp(float3);
-short3 __ovld __cnfn convert_short3_sat_rtp(float3);
-short3 __ovld __cnfn convert_short3_rtn(float3);
-short3 __ovld __cnfn convert_short3_sat_rtn(float3);
-short3 __ovld __cnfn convert_short3(float3);
-short3 __ovld __cnfn convert_short3_sat(float3);
-ushort3 __ovld __cnfn convert_ushort3_rte(char3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(char3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(char3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(char3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(char3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(char3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(char3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(char3);
-ushort3 __ovld __cnfn convert_ushort3(char3);
-ushort3 __ovld __cnfn convert_ushort3_sat(char3);
-ushort3 __ovld __cnfn convert_ushort3_rte(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(uchar3);
-ushort3 __ovld __cnfn convert_ushort3(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_sat(uchar3);
-ushort3 __ovld __cnfn convert_ushort3_rte(short3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(short3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(short3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(short3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(short3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(short3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(short3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(short3);
-ushort3 __ovld __cnfn convert_ushort3(short3);
-ushort3 __ovld __cnfn convert_ushort3_sat(short3);
-ushort3 __ovld __cnfn convert_ushort3_rte(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(ushort3);
-ushort3 __ovld __cnfn convert_ushort3(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_sat(ushort3);
-ushort3 __ovld __cnfn convert_ushort3_rte(int3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(int3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(int3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(int3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(int3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(int3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(int3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(int3);
-ushort3 __ovld __cnfn convert_ushort3(int3);
-ushort3 __ovld __cnfn convert_ushort3_sat(int3);
-ushort3 __ovld __cnfn convert_ushort3_rte(uint3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(uint3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(uint3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(uint3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(uint3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(uint3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(uint3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(uint3);
-ushort3 __ovld __cnfn convert_ushort3(uint3);
-ushort3 __ovld __cnfn convert_ushort3_sat(uint3);
-ushort3 __ovld __cnfn convert_ushort3_rte(long3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(long3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(long3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(long3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(long3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(long3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(long3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(long3);
-ushort3 __ovld __cnfn convert_ushort3(long3);
-ushort3 __ovld __cnfn convert_ushort3_sat(long3);
-ushort3 __ovld __cnfn convert_ushort3_rte(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(ulong3);
-ushort3 __ovld __cnfn convert_ushort3(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_sat(ulong3);
-ushort3 __ovld __cnfn convert_ushort3_rte(float3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(float3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(float3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(float3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(float3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(float3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(float3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(float3);
-ushort3 __ovld __cnfn convert_ushort3(float3);
-ushort3 __ovld __cnfn convert_ushort3_sat(float3);
-int3 __ovld __cnfn convert_int3_rte(char3);
-int3 __ovld __cnfn convert_int3_sat_rte(char3);
-int3 __ovld __cnfn convert_int3_rtz(char3);
-int3 __ovld __cnfn convert_int3_sat_rtz(char3);
-int3 __ovld __cnfn convert_int3_rtp(char3);
-int3 __ovld __cnfn convert_int3_sat_rtp(char3);
-int3 __ovld __cnfn convert_int3_rtn(char3);
-int3 __ovld __cnfn convert_int3_sat_rtn(char3);
-int3 __ovld __cnfn convert_int3(char3);
-int3 __ovld __cnfn convert_int3_sat(char3);
-int3 __ovld __cnfn convert_int3_rte(uchar3);
-int3 __ovld __cnfn convert_int3_sat_rte(uchar3);
-int3 __ovld __cnfn convert_int3_rtz(uchar3);
-int3 __ovld __cnfn convert_int3_sat_rtz(uchar3);
-int3 __ovld __cnfn convert_int3_rtp(uchar3);
-int3 __ovld __cnfn convert_int3_sat_rtp(uchar3);
-int3 __ovld __cnfn convert_int3_rtn(uchar3);
-int3 __ovld __cnfn convert_int3_sat_rtn(uchar3);
-int3 __ovld __cnfn convert_int3(uchar3);
-int3 __ovld __cnfn convert_int3_sat(uchar3);
-int3 __ovld __cnfn convert_int3_rte(short3);
-int3 __ovld __cnfn convert_int3_sat_rte(short3);
-int3 __ovld __cnfn convert_int3_rtz(short3);
-int3 __ovld __cnfn convert_int3_sat_rtz(short3);
-int3 __ovld __cnfn convert_int3_rtp(short3);
-int3 __ovld __cnfn convert_int3_sat_rtp(short3);
-int3 __ovld __cnfn convert_int3_rtn(short3);
-int3 __ovld __cnfn convert_int3_sat_rtn(short3);
-int3 __ovld __cnfn convert_int3(short3);
-int3 __ovld __cnfn convert_int3_sat(short3);
-int3 __ovld __cnfn convert_int3_rte(ushort3);
-int3 __ovld __cnfn convert_int3_sat_rte(ushort3);
-int3 __ovld __cnfn convert_int3_rtz(ushort3);
-int3 __ovld __cnfn convert_int3_sat_rtz(ushort3);
-int3 __ovld __cnfn convert_int3_rtp(ushort3);
-int3 __ovld __cnfn convert_int3_sat_rtp(ushort3);
-int3 __ovld __cnfn convert_int3_rtn(ushort3);
-int3 __ovld __cnfn convert_int3_sat_rtn(ushort3);
-int3 __ovld __cnfn convert_int3(ushort3);
-int3 __ovld __cnfn convert_int3_sat(ushort3);
-int3 __ovld __cnfn convert_int3_rte(int3);
-int3 __ovld __cnfn convert_int3_sat_rte(int3);
-int3 __ovld __cnfn convert_int3_rtz(int3);
-int3 __ovld __cnfn convert_int3_sat_rtz(int3);
-int3 __ovld __cnfn convert_int3_rtp(int3);
-int3 __ovld __cnfn convert_int3_sat_rtp(int3);
-int3 __ovld __cnfn convert_int3_rtn(int3);
-int3 __ovld __cnfn convert_int3_sat_rtn(int3);
-int3 __ovld __cnfn convert_int3(int3);
-int3 __ovld __cnfn convert_int3_sat(int3);
-int3 __ovld __cnfn convert_int3_rte(uint3);
-int3 __ovld __cnfn convert_int3_sat_rte(uint3);
-int3 __ovld __cnfn convert_int3_rtz(uint3);
-int3 __ovld __cnfn convert_int3_sat_rtz(uint3);
-int3 __ovld __cnfn convert_int3_rtp(uint3);
-int3 __ovld __cnfn convert_int3_sat_rtp(uint3);
-int3 __ovld __cnfn convert_int3_rtn(uint3);
-int3 __ovld __cnfn convert_int3_sat_rtn(uint3);
-int3 __ovld __cnfn convert_int3(uint3);
-int3 __ovld __cnfn convert_int3_sat(uint3);
-int3 __ovld __cnfn convert_int3_rte(long3);
-int3 __ovld __cnfn convert_int3_sat_rte(long3);
-int3 __ovld __cnfn convert_int3_rtz(long3);
-int3 __ovld __cnfn convert_int3_sat_rtz(long3);
-int3 __ovld __cnfn convert_int3_rtp(long3);
-int3 __ovld __cnfn convert_int3_sat_rtp(long3);
-int3 __ovld __cnfn convert_int3_rtn(long3);
-int3 __ovld __cnfn convert_int3_sat_rtn(long3);
-int3 __ovld __cnfn convert_int3(long3);
-int3 __ovld __cnfn convert_int3_sat(long3);
-int3 __ovld __cnfn convert_int3_rte(ulong3);
-int3 __ovld __cnfn convert_int3_sat_rte(ulong3);
-int3 __ovld __cnfn convert_int3_rtz(ulong3);
-int3 __ovld __cnfn convert_int3_sat_rtz(ulong3);
-int3 __ovld __cnfn convert_int3_rtp(ulong3);
-int3 __ovld __cnfn convert_int3_sat_rtp(ulong3);
-int3 __ovld __cnfn convert_int3_rtn(ulong3);
-int3 __ovld __cnfn convert_int3_sat_rtn(ulong3);
-int3 __ovld __cnfn convert_int3(ulong3);
-int3 __ovld __cnfn convert_int3_sat(ulong3);
-int3 __ovld __cnfn convert_int3_rte(float3);
-int3 __ovld __cnfn convert_int3_sat_rte(float3);
-int3 __ovld __cnfn convert_int3_rtz(float3);
-int3 __ovld __cnfn convert_int3_sat_rtz(float3);
-int3 __ovld __cnfn convert_int3_rtp(float3);
-int3 __ovld __cnfn convert_int3_sat_rtp(float3);
-int3 __ovld __cnfn convert_int3_rtn(float3);
-int3 __ovld __cnfn convert_int3_sat_rtn(float3);
-int3 __ovld __cnfn convert_int3(float3);
-int3 __ovld __cnfn convert_int3_sat(float3);
-uint3 __ovld __cnfn convert_uint3_rte(char3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(char3);
-uint3 __ovld __cnfn convert_uint3_rtz(char3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(char3);
-uint3 __ovld __cnfn convert_uint3_rtp(char3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(char3);
-uint3 __ovld __cnfn convert_uint3_rtn(char3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(char3);
-uint3 __ovld __cnfn convert_uint3(char3);
-uint3 __ovld __cnfn convert_uint3_sat(char3);
-uint3 __ovld __cnfn convert_uint3_rte(uchar3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(uchar3);
-uint3 __ovld __cnfn convert_uint3_rtz(uchar3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(uchar3);
-uint3 __ovld __cnfn convert_uint3_rtp(uchar3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(uchar3);
-uint3 __ovld __cnfn convert_uint3_rtn(uchar3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(uchar3);
-uint3 __ovld __cnfn convert_uint3(uchar3);
-uint3 __ovld __cnfn convert_uint3_sat(uchar3);
-uint3 __ovld __cnfn convert_uint3_rte(short3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(short3);
-uint3 __ovld __cnfn convert_uint3_rtz(short3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(short3);
-uint3 __ovld __cnfn convert_uint3_rtp(short3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(short3);
-uint3 __ovld __cnfn convert_uint3_rtn(short3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(short3);
-uint3 __ovld __cnfn convert_uint3(short3);
-uint3 __ovld __cnfn convert_uint3_sat(short3);
-uint3 __ovld __cnfn convert_uint3_rte(ushort3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(ushort3);
-uint3 __ovld __cnfn convert_uint3_rtz(ushort3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(ushort3);
-uint3 __ovld __cnfn convert_uint3_rtp(ushort3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(ushort3);
-uint3 __ovld __cnfn convert_uint3_rtn(ushort3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(ushort3);
-uint3 __ovld __cnfn convert_uint3(ushort3);
-uint3 __ovld __cnfn convert_uint3_sat(ushort3);
-uint3 __ovld __cnfn convert_uint3_rte(int3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(int3);
-uint3 __ovld __cnfn convert_uint3_rtz(int3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(int3);
-uint3 __ovld __cnfn convert_uint3_rtp(int3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(int3);
-uint3 __ovld __cnfn convert_uint3_rtn(int3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(int3);
-uint3 __ovld __cnfn convert_uint3(int3);
-uint3 __ovld __cnfn convert_uint3_sat(int3);
-uint3 __ovld __cnfn convert_uint3_rte(uint3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(uint3);
-uint3 __ovld __cnfn convert_uint3_rtz(uint3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(uint3);
-uint3 __ovld __cnfn convert_uint3_rtp(uint3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(uint3);
-uint3 __ovld __cnfn convert_uint3_rtn(uint3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(uint3);
-uint3 __ovld __cnfn convert_uint3(uint3);
-uint3 __ovld __cnfn convert_uint3_sat(uint3);
-uint3 __ovld __cnfn convert_uint3_rte(long3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(long3);
-uint3 __ovld __cnfn convert_uint3_rtz(long3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(long3);
-uint3 __ovld __cnfn convert_uint3_rtp(long3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(long3);
-uint3 __ovld __cnfn convert_uint3_rtn(long3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(long3);
-uint3 __ovld __cnfn convert_uint3(long3);
-uint3 __ovld __cnfn convert_uint3_sat(long3);
-uint3 __ovld __cnfn convert_uint3_rte(ulong3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(ulong3);
-uint3 __ovld __cnfn convert_uint3_rtz(ulong3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(ulong3);
-uint3 __ovld __cnfn convert_uint3_rtp(ulong3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(ulong3);
-uint3 __ovld __cnfn convert_uint3_rtn(ulong3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(ulong3);
-uint3 __ovld __cnfn convert_uint3(ulong3);
-uint3 __ovld __cnfn convert_uint3_sat(ulong3);
-uint3 __ovld __cnfn convert_uint3_rte(float3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(float3);
-uint3 __ovld __cnfn convert_uint3_rtz(float3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(float3);
-uint3 __ovld __cnfn convert_uint3_rtp(float3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(float3);
-uint3 __ovld __cnfn convert_uint3_rtn(float3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(float3);
-uint3 __ovld __cnfn convert_uint3(float3);
-uint3 __ovld __cnfn convert_uint3_sat(float3);
-long3 __ovld __cnfn convert_long3_rte(char3);
-long3 __ovld __cnfn convert_long3_sat_rte(char3);
-long3 __ovld __cnfn convert_long3_rtz(char3);
-long3 __ovld __cnfn convert_long3_sat_rtz(char3);
-long3 __ovld __cnfn convert_long3_rtp(char3);
-long3 __ovld __cnfn convert_long3_sat_rtp(char3);
-long3 __ovld __cnfn convert_long3_rtn(char3);
-long3 __ovld __cnfn convert_long3_sat_rtn(char3);
-long3 __ovld __cnfn convert_long3(char3);
-long3 __ovld __cnfn convert_long3_sat(char3);
-long3 __ovld __cnfn convert_long3_rte(uchar3);
-long3 __ovld __cnfn convert_long3_sat_rte(uchar3);
-long3 __ovld __cnfn convert_long3_rtz(uchar3);
-long3 __ovld __cnfn convert_long3_sat_rtz(uchar3);
-long3 __ovld __cnfn convert_long3_rtp(uchar3);
-long3 __ovld __cnfn convert_long3_sat_rtp(uchar3);
-long3 __ovld __cnfn convert_long3_rtn(uchar3);
-long3 __ovld __cnfn convert_long3_sat_rtn(uchar3);
-long3 __ovld __cnfn convert_long3(uchar3);
-long3 __ovld __cnfn convert_long3_sat(uchar3);
-long3 __ovld __cnfn convert_long3_rte(short3);
-long3 __ovld __cnfn convert_long3_sat_rte(short3);
-long3 __ovld __cnfn convert_long3_rtz(short3);
-long3 __ovld __cnfn convert_long3_sat_rtz(short3);
-long3 __ovld __cnfn convert_long3_rtp(short3);
-long3 __ovld __cnfn convert_long3_sat_rtp(short3);
-long3 __ovld __cnfn convert_long3_rtn(short3);
-long3 __ovld __cnfn convert_long3_sat_rtn(short3);
-long3 __ovld __cnfn convert_long3(short3);
-long3 __ovld __cnfn convert_long3_sat(short3);
-long3 __ovld __cnfn convert_long3_rte(ushort3);
-long3 __ovld __cnfn convert_long3_sat_rte(ushort3);
-long3 __ovld __cnfn convert_long3_rtz(ushort3);
-long3 __ovld __cnfn convert_long3_sat_rtz(ushort3);
-long3 __ovld __cnfn convert_long3_rtp(ushort3);
-long3 __ovld __cnfn convert_long3_sat_rtp(ushort3);
-long3 __ovld __cnfn convert_long3_rtn(ushort3);
-long3 __ovld __cnfn convert_long3_sat_rtn(ushort3);
-long3 __ovld __cnfn convert_long3(ushort3);
-long3 __ovld __cnfn convert_long3_sat(ushort3);
-long3 __ovld __cnfn convert_long3_rte(int3);
-long3 __ovld __cnfn convert_long3_sat_rte(int3);
-long3 __ovld __cnfn convert_long3_rtz(int3);
-long3 __ovld __cnfn convert_long3_sat_rtz(int3);
-long3 __ovld __cnfn convert_long3_rtp(int3);
-long3 __ovld __cnfn convert_long3_sat_rtp(int3);
-long3 __ovld __cnfn convert_long3_rtn(int3);
-long3 __ovld __cnfn convert_long3_sat_rtn(int3);
-long3 __ovld __cnfn convert_long3(int3);
-long3 __ovld __cnfn convert_long3_sat(int3);
-long3 __ovld __cnfn convert_long3_rte(uint3);
-long3 __ovld __cnfn convert_long3_sat_rte(uint3);
-long3 __ovld __cnfn convert_long3_rtz(uint3);
-long3 __ovld __cnfn convert_long3_sat_rtz(uint3);
-long3 __ovld __cnfn convert_long3_rtp(uint3);
-long3 __ovld __cnfn convert_long3_sat_rtp(uint3);
-long3 __ovld __cnfn convert_long3_rtn(uint3);
-long3 __ovld __cnfn convert_long3_sat_rtn(uint3);
-long3 __ovld __cnfn convert_long3(uint3);
-long3 __ovld __cnfn convert_long3_sat(uint3);
-long3 __ovld __cnfn convert_long3_rte(long3);
-long3 __ovld __cnfn convert_long3_sat_rte(long3);
-long3 __ovld __cnfn convert_long3_rtz(long3);
-long3 __ovld __cnfn convert_long3_sat_rtz(long3);
-long3 __ovld __cnfn convert_long3_rtp(long3);
-long3 __ovld __cnfn convert_long3_sat_rtp(long3);
-long3 __ovld __cnfn convert_long3_rtn(long3);
-long3 __ovld __cnfn convert_long3_sat_rtn(long3);
-long3 __ovld __cnfn convert_long3(long3);
-long3 __ovld __cnfn convert_long3_sat(long3);
-long3 __ovld __cnfn convert_long3_rte(ulong3);
-long3 __ovld __cnfn convert_long3_sat_rte(ulong3);
-long3 __ovld __cnfn convert_long3_rtz(ulong3);
-long3 __ovld __cnfn convert_long3_sat_rtz(ulong3);
-long3 __ovld __cnfn convert_long3_rtp(ulong3);
-long3 __ovld __cnfn convert_long3_sat_rtp(ulong3);
-long3 __ovld __cnfn convert_long3_rtn(ulong3);
-long3 __ovld __cnfn convert_long3_sat_rtn(ulong3);
-long3 __ovld __cnfn convert_long3(ulong3);
-long3 __ovld __cnfn convert_long3_sat(ulong3);
-long3 __ovld __cnfn convert_long3_rte(float3);
-long3 __ovld __cnfn convert_long3_sat_rte(float3);
-long3 __ovld __cnfn convert_long3_rtz(float3);
-long3 __ovld __cnfn convert_long3_sat_rtz(float3);
-long3 __ovld __cnfn convert_long3_rtp(float3);
-long3 __ovld __cnfn convert_long3_sat_rtp(float3);
-long3 __ovld __cnfn convert_long3_rtn(float3);
-long3 __ovld __cnfn convert_long3_sat_rtn(float3);
-long3 __ovld __cnfn convert_long3(float3);
-long3 __ovld __cnfn convert_long3_sat(float3);
-ulong3 __ovld __cnfn convert_ulong3_rte(char3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(char3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(char3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(char3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(char3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(char3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(char3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(char3);
-ulong3 __ovld __cnfn convert_ulong3(char3);
-ulong3 __ovld __cnfn convert_ulong3_sat(char3);
-ulong3 __ovld __cnfn convert_ulong3_rte(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(uchar3);
-ulong3 __ovld __cnfn convert_ulong3(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_sat(uchar3);
-ulong3 __ovld __cnfn convert_ulong3_rte(short3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(short3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(short3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(short3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(short3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(short3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(short3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(short3);
-ulong3 __ovld __cnfn convert_ulong3(short3);
-ulong3 __ovld __cnfn convert_ulong3_sat(short3);
-ulong3 __ovld __cnfn convert_ulong3_rte(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(ushort3);
-ulong3 __ovld __cnfn convert_ulong3(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_sat(ushort3);
-ulong3 __ovld __cnfn convert_ulong3_rte(int3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(int3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(int3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(int3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(int3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(int3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(int3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(int3);
-ulong3 __ovld __cnfn convert_ulong3(int3);
-ulong3 __ovld __cnfn convert_ulong3_sat(int3);
-ulong3 __ovld __cnfn convert_ulong3_rte(uint3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(uint3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(uint3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(uint3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(uint3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(uint3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(uint3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(uint3);
-ulong3 __ovld __cnfn convert_ulong3(uint3);
-ulong3 __ovld __cnfn convert_ulong3_sat(uint3);
-ulong3 __ovld __cnfn convert_ulong3_rte(long3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(long3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(long3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(long3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(long3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(long3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(long3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(long3);
-ulong3 __ovld __cnfn convert_ulong3(long3);
-ulong3 __ovld __cnfn convert_ulong3_sat(long3);
-ulong3 __ovld __cnfn convert_ulong3_rte(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(ulong3);
-ulong3 __ovld __cnfn convert_ulong3(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_sat(ulong3);
-ulong3 __ovld __cnfn convert_ulong3_rte(float3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(float3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(float3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(float3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(float3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(float3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(float3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(float3);
-ulong3 __ovld __cnfn convert_ulong3(float3);
-ulong3 __ovld __cnfn convert_ulong3_sat(float3);
-float3 __ovld __cnfn convert_float3_rte(char3);
-float3 __ovld __cnfn convert_float3_rtz(char3);
-float3 __ovld __cnfn convert_float3_rtp(char3);
-float3 __ovld __cnfn convert_float3_rtn(char3);
-float3 __ovld __cnfn convert_float3(char3);
-float3 __ovld __cnfn convert_float3_rte(uchar3);
-float3 __ovld __cnfn convert_float3_rtz(uchar3);
-float3 __ovld __cnfn convert_float3_rtp(uchar3);
-float3 __ovld __cnfn convert_float3_rtn(uchar3);
-float3 __ovld __cnfn convert_float3(uchar3);
-float3 __ovld __cnfn convert_float3_rte(short3);
-float3 __ovld __cnfn convert_float3_rtz(short3);
-float3 __ovld __cnfn convert_float3_rtp(short3);
-float3 __ovld __cnfn convert_float3_rtn(short3);
-float3 __ovld __cnfn convert_float3(short3);
-float3 __ovld __cnfn convert_float3_rte(ushort3);
-float3 __ovld __cnfn convert_float3_rtz(ushort3);
-float3 __ovld __cnfn convert_float3_rtp(ushort3);
-float3 __ovld __cnfn convert_float3_rtn(ushort3);
-float3 __ovld __cnfn convert_float3(ushort3);
-float3 __ovld __cnfn convert_float3_rte(int3);
-float3 __ovld __cnfn convert_float3_rtz(int3);
-float3 __ovld __cnfn convert_float3_rtp(int3);
-float3 __ovld __cnfn convert_float3_rtn(int3);
-float3 __ovld __cnfn convert_float3(int3);
-float3 __ovld __cnfn convert_float3_rte(uint3);
-float3 __ovld __cnfn convert_float3_rtz(uint3);
-float3 __ovld __cnfn convert_float3_rtp(uint3);
-float3 __ovld __cnfn convert_float3_rtn(uint3);
-float3 __ovld __cnfn convert_float3(uint3);
-float3 __ovld __cnfn convert_float3_rte(long3);
-float3 __ovld __cnfn convert_float3_rtz(long3);
-float3 __ovld __cnfn convert_float3_rtp(long3);
-float3 __ovld __cnfn convert_float3_rtn(long3);
-float3 __ovld __cnfn convert_float3(long3);
-float3 __ovld __cnfn convert_float3_rte(ulong3);
-float3 __ovld __cnfn convert_float3_rtz(ulong3);
-float3 __ovld __cnfn convert_float3_rtp(ulong3);
-float3 __ovld __cnfn convert_float3_rtn(ulong3);
-float3 __ovld __cnfn convert_float3(ulong3);
-float3 __ovld __cnfn convert_float3_rte(float3);
-float3 __ovld __cnfn convert_float3_rtz(float3);
-float3 __ovld __cnfn convert_float3_rtp(float3);
-float3 __ovld __cnfn convert_float3_rtn(float3);
-float3 __ovld __cnfn convert_float3(float3);
-char4 __ovld __cnfn convert_char4_rte(char4);
-char4 __ovld __cnfn convert_char4_sat_rte(char4);
-char4 __ovld __cnfn convert_char4_rtz(char4);
-char4 __ovld __cnfn convert_char4_sat_rtz(char4);
-char4 __ovld __cnfn convert_char4_rtp(char4);
-char4 __ovld __cnfn convert_char4_sat_rtp(char4);
-char4 __ovld __cnfn convert_char4_rtn(char4);
-char4 __ovld __cnfn convert_char4_sat_rtn(char4);
-char4 __ovld __cnfn convert_char4(char4);
-char4 __ovld __cnfn convert_char4_sat(char4);
-char4 __ovld __cnfn convert_char4_rte(uchar4);
-char4 __ovld __cnfn convert_char4_sat_rte(uchar4);
-char4 __ovld __cnfn convert_char4_rtz(uchar4);
-char4 __ovld __cnfn convert_char4_sat_rtz(uchar4);
-char4 __ovld __cnfn convert_char4_rtp(uchar4);
-char4 __ovld __cnfn convert_char4_sat_rtp(uchar4);
-char4 __ovld __cnfn convert_char4_rtn(uchar4);
-char4 __ovld __cnfn convert_char4_sat_rtn(uchar4);
-char4 __ovld __cnfn convert_char4(uchar4);
-char4 __ovld __cnfn convert_char4_sat(uchar4);
-char4 __ovld __cnfn convert_char4_rte(short4);
-char4 __ovld __cnfn convert_char4_sat_rte(short4);
-char4 __ovld __cnfn convert_char4_rtz(short4);
-char4 __ovld __cnfn convert_char4_sat_rtz(short4);
-char4 __ovld __cnfn convert_char4_rtp(short4);
-char4 __ovld __cnfn convert_char4_sat_rtp(short4);
-char4 __ovld __cnfn convert_char4_rtn(short4);
-char4 __ovld __cnfn convert_char4_sat_rtn(short4);
-char4 __ovld __cnfn convert_char4(short4);
-char4 __ovld __cnfn convert_char4_sat(short4);
-char4 __ovld __cnfn convert_char4_rte(ushort4);
-char4 __ovld __cnfn convert_char4_sat_rte(ushort4);
-char4 __ovld __cnfn convert_char4_rtz(ushort4);
-char4 __ovld __cnfn convert_char4_sat_rtz(ushort4);
-char4 __ovld __cnfn convert_char4_rtp(ushort4);
-char4 __ovld __cnfn convert_char4_sat_rtp(ushort4);
-char4 __ovld __cnfn convert_char4_rtn(ushort4);
-char4 __ovld __cnfn convert_char4_sat_rtn(ushort4);
-char4 __ovld __cnfn convert_char4(ushort4);
-char4 __ovld __cnfn convert_char4_sat(ushort4);
-char4 __ovld __cnfn convert_char4_rte(int4);
-char4 __ovld __cnfn convert_char4_sat_rte(int4);
-char4 __ovld __cnfn convert_char4_rtz(int4);
-char4 __ovld __cnfn convert_char4_sat_rtz(int4);
-char4 __ovld __cnfn convert_char4_rtp(int4);
-char4 __ovld __cnfn convert_char4_sat_rtp(int4);
-char4 __ovld __cnfn convert_char4_rtn(int4);
-char4 __ovld __cnfn convert_char4_sat_rtn(int4);
-char4 __ovld __cnfn convert_char4(int4);
-char4 __ovld __cnfn convert_char4_sat(int4);
-char4 __ovld __cnfn convert_char4_rte(uint4);
-char4 __ovld __cnfn convert_char4_sat_rte(uint4);
-char4 __ovld __cnfn convert_char4_rtz(uint4);
-char4 __ovld __cnfn convert_char4_sat_rtz(uint4);
-char4 __ovld __cnfn convert_char4_rtp(uint4);
-char4 __ovld __cnfn convert_char4_sat_rtp(uint4);
-char4 __ovld __cnfn convert_char4_rtn(uint4);
-char4 __ovld __cnfn convert_char4_sat_rtn(uint4);
-char4 __ovld __cnfn convert_char4(uint4);
-char4 __ovld __cnfn convert_char4_sat(uint4);
-char4 __ovld __cnfn convert_char4_rte(long4);
-char4 __ovld __cnfn convert_char4_sat_rte(long4);
-char4 __ovld __cnfn convert_char4_rtz(long4);
-char4 __ovld __cnfn convert_char4_sat_rtz(long4);
-char4 __ovld __cnfn convert_char4_rtp(long4);
-char4 __ovld __cnfn convert_char4_sat_rtp(long4);
-char4 __ovld __cnfn convert_char4_rtn(long4);
-char4 __ovld __cnfn convert_char4_sat_rtn(long4);
-char4 __ovld __cnfn convert_char4(long4);
-char4 __ovld __cnfn convert_char4_sat(long4);
-char4 __ovld __cnfn convert_char4_rte(ulong4);
-char4 __ovld __cnfn convert_char4_sat_rte(ulong4);
-char4 __ovld __cnfn convert_char4_rtz(ulong4);
-char4 __ovld __cnfn convert_char4_sat_rtz(ulong4);
-char4 __ovld __cnfn convert_char4_rtp(ulong4);
-char4 __ovld __cnfn convert_char4_sat_rtp(ulong4);
-char4 __ovld __cnfn convert_char4_rtn(ulong4);
-char4 __ovld __cnfn convert_char4_sat_rtn(ulong4);
-char4 __ovld __cnfn convert_char4(ulong4);
-char4 __ovld __cnfn convert_char4_sat(ulong4);
-char4 __ovld __cnfn convert_char4_rte(float4);
-char4 __ovld __cnfn convert_char4_sat_rte(float4);
-char4 __ovld __cnfn convert_char4_rtz(float4);
-char4 __ovld __cnfn convert_char4_sat_rtz(float4);
-char4 __ovld __cnfn convert_char4_rtp(float4);
-char4 __ovld __cnfn convert_char4_sat_rtp(float4);
-char4 __ovld __cnfn convert_char4_rtn(float4);
-char4 __ovld __cnfn convert_char4_sat_rtn(float4);
-char4 __ovld __cnfn convert_char4(float4);
-char4 __ovld __cnfn convert_char4_sat(float4);
-uchar4 __ovld __cnfn convert_uchar4_rte(char4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(char4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(char4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(char4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(char4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(char4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(char4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(char4);
-uchar4 __ovld __cnfn convert_uchar4(char4);
-uchar4 __ovld __cnfn convert_uchar4_sat(char4);
-uchar4 __ovld __cnfn convert_uchar4_rte(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(uchar4);
-uchar4 __ovld __cnfn convert_uchar4(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_sat(uchar4);
-uchar4 __ovld __cnfn convert_uchar4_rte(short4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(short4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(short4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(short4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(short4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(short4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(short4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(short4);
-uchar4 __ovld __cnfn convert_uchar4(short4);
-uchar4 __ovld __cnfn convert_uchar4_sat(short4);
-uchar4 __ovld __cnfn convert_uchar4_rte(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(ushort4);
-uchar4 __ovld __cnfn convert_uchar4(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_sat(ushort4);
-uchar4 __ovld __cnfn convert_uchar4_rte(int4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(int4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(int4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(int4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(int4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(int4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(int4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(int4);
-uchar4 __ovld __cnfn convert_uchar4(int4);
-uchar4 __ovld __cnfn convert_uchar4_sat(int4);
-uchar4 __ovld __cnfn convert_uchar4_rte(uint4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(uint4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(uint4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(uint4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(uint4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(uint4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(uint4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(uint4);
-uchar4 __ovld __cnfn convert_uchar4(uint4);
-uchar4 __ovld __cnfn convert_uchar4_sat(uint4);
-uchar4 __ovld __cnfn convert_uchar4_rte(long4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(long4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(long4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(long4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(long4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(long4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(long4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(long4);
-uchar4 __ovld __cnfn convert_uchar4(long4);
-uchar4 __ovld __cnfn convert_uchar4_sat(long4);
-uchar4 __ovld __cnfn convert_uchar4_rte(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(ulong4);
-uchar4 __ovld __cnfn convert_uchar4(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_sat(ulong4);
-uchar4 __ovld __cnfn convert_uchar4_rte(float4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(float4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(float4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(float4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(float4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(float4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(float4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(float4);
-uchar4 __ovld __cnfn convert_uchar4(float4);
-uchar4 __ovld __cnfn convert_uchar4_sat(float4);
-short4 __ovld __cnfn convert_short4_rte(char4);
-short4 __ovld __cnfn convert_short4_sat_rte(char4);
-short4 __ovld __cnfn convert_short4_rtz(char4);
-short4 __ovld __cnfn convert_short4_sat_rtz(char4);
-short4 __ovld __cnfn convert_short4_rtp(char4);
-short4 __ovld __cnfn convert_short4_sat_rtp(char4);
-short4 __ovld __cnfn convert_short4_rtn(char4);
-short4 __ovld __cnfn convert_short4_sat_rtn(char4);
-short4 __ovld __cnfn convert_short4(char4);
-short4 __ovld __cnfn convert_short4_sat(char4);
-short4 __ovld __cnfn convert_short4_rte(uchar4);
-short4 __ovld __cnfn convert_short4_sat_rte(uchar4);
-short4 __ovld __cnfn convert_short4_rtz(uchar4);
-short4 __ovld __cnfn convert_short4_sat_rtz(uchar4);
-short4 __ovld __cnfn convert_short4_rtp(uchar4);
-short4 __ovld __cnfn convert_short4_sat_rtp(uchar4);
-short4 __ovld __cnfn convert_short4_rtn(uchar4);
-short4 __ovld __cnfn convert_short4_sat_rtn(uchar4);
-short4 __ovld __cnfn convert_short4(uchar4);
-short4 __ovld __cnfn convert_short4_sat(uchar4);
-short4 __ovld __cnfn convert_short4_rte(short4);
-short4 __ovld __cnfn convert_short4_sat_rte(short4);
-short4 __ovld __cnfn convert_short4_rtz(short4);
-short4 __ovld __cnfn convert_short4_sat_rtz(short4);
-short4 __ovld __cnfn convert_short4_rtp(short4);
-short4 __ovld __cnfn convert_short4_sat_rtp(short4);
-short4 __ovld __cnfn convert_short4_rtn(short4);
-short4 __ovld __cnfn convert_short4_sat_rtn(short4);
-short4 __ovld __cnfn convert_short4(short4);
-short4 __ovld __cnfn convert_short4_sat(short4);
-short4 __ovld __cnfn convert_short4_rte(ushort4);
-short4 __ovld __cnfn convert_short4_sat_rte(ushort4);
-short4 __ovld __cnfn convert_short4_rtz(ushort4);
-short4 __ovld __cnfn convert_short4_sat_rtz(ushort4);
-short4 __ovld __cnfn convert_short4_rtp(ushort4);
-short4 __ovld __cnfn convert_short4_sat_rtp(ushort4);
-short4 __ovld __cnfn convert_short4_rtn(ushort4);
-short4 __ovld __cnfn convert_short4_sat_rtn(ushort4);
-short4 __ovld __cnfn convert_short4(ushort4);
-short4 __ovld __cnfn convert_short4_sat(ushort4);
-short4 __ovld __cnfn convert_short4_rte(int4);
-short4 __ovld __cnfn convert_short4_sat_rte(int4);
-short4 __ovld __cnfn convert_short4_rtz(int4);
-short4 __ovld __cnfn convert_short4_sat_rtz(int4);
-short4 __ovld __cnfn convert_short4_rtp(int4);
-short4 __ovld __cnfn convert_short4_sat_rtp(int4);
-short4 __ovld __cnfn convert_short4_rtn(int4);
-short4 __ovld __cnfn convert_short4_sat_rtn(int4);
-short4 __ovld __cnfn convert_short4(int4);
-short4 __ovld __cnfn convert_short4_sat(int4);
-short4 __ovld __cnfn convert_short4_rte(uint4);
-short4 __ovld __cnfn convert_short4_sat_rte(uint4);
-short4 __ovld __cnfn convert_short4_rtz(uint4);
-short4 __ovld __cnfn convert_short4_sat_rtz(uint4);
-short4 __ovld __cnfn convert_short4_rtp(uint4);
-short4 __ovld __cnfn convert_short4_sat_rtp(uint4);
-short4 __ovld __cnfn convert_short4_rtn(uint4);
-short4 __ovld __cnfn convert_short4_sat_rtn(uint4);
-short4 __ovld __cnfn convert_short4(uint4);
-short4 __ovld __cnfn convert_short4_sat(uint4);
-short4 __ovld __cnfn convert_short4_rte(long4);
-short4 __ovld __cnfn convert_short4_sat_rte(long4);
-short4 __ovld __cnfn convert_short4_rtz(long4);
-short4 __ovld __cnfn convert_short4_sat_rtz(long4);
-short4 __ovld __cnfn convert_short4_rtp(long4);
-short4 __ovld __cnfn convert_short4_sat_rtp(long4);
-short4 __ovld __cnfn convert_short4_rtn(long4);
-short4 __ovld __cnfn convert_short4_sat_rtn(long4);
-short4 __ovld __cnfn convert_short4(long4);
-short4 __ovld __cnfn convert_short4_sat(long4);
-short4 __ovld __cnfn convert_short4_rte(ulong4);
-short4 __ovld __cnfn convert_short4_sat_rte(ulong4);
-short4 __ovld __cnfn convert_short4_rtz(ulong4);
-short4 __ovld __cnfn convert_short4_sat_rtz(ulong4);
-short4 __ovld __cnfn convert_short4_rtp(ulong4);
-short4 __ovld __cnfn convert_short4_sat_rtp(ulong4);
-short4 __ovld __cnfn convert_short4_rtn(ulong4);
-short4 __ovld __cnfn convert_short4_sat_rtn(ulong4);
-short4 __ovld __cnfn convert_short4(ulong4);
-short4 __ovld __cnfn convert_short4_sat(ulong4);
-short4 __ovld __cnfn convert_short4_rte(float4);
-short4 __ovld __cnfn convert_short4_sat_rte(float4);
-short4 __ovld __cnfn convert_short4_rtz(float4);
-short4 __ovld __cnfn convert_short4_sat_rtz(float4);
-short4 __ovld __cnfn convert_short4_rtp(float4);
-short4 __ovld __cnfn convert_short4_sat_rtp(float4);
-short4 __ovld __cnfn convert_short4_rtn(float4);
-short4 __ovld __cnfn convert_short4_sat_rtn(float4);
-short4 __ovld __cnfn convert_short4(float4);
-short4 __ovld __cnfn convert_short4_sat(float4);
-ushort4 __ovld __cnfn convert_ushort4_rte(char4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(char4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(char4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(char4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(char4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(char4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(char4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(char4);
-ushort4 __ovld __cnfn convert_ushort4(char4);
-ushort4 __ovld __cnfn convert_ushort4_sat(char4);
-ushort4 __ovld __cnfn convert_ushort4_rte(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(uchar4);
-ushort4 __ovld __cnfn convert_ushort4(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_sat(uchar4);
-ushort4 __ovld __cnfn convert_ushort4_rte(short4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(short4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(short4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(short4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(short4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(short4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(short4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(short4);
-ushort4 __ovld __cnfn convert_ushort4(short4);
-ushort4 __ovld __cnfn convert_ushort4_sat(short4);
-ushort4 __ovld __cnfn convert_ushort4_rte(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(ushort4);
-ushort4 __ovld __cnfn convert_ushort4(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_sat(ushort4);
-ushort4 __ovld __cnfn convert_ushort4_rte(int4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(int4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(int4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(int4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(int4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(int4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(int4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(int4);
-ushort4 __ovld __cnfn convert_ushort4(int4);
-ushort4 __ovld __cnfn convert_ushort4_sat(int4);
-ushort4 __ovld __cnfn convert_ushort4_rte(uint4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(uint4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(uint4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(uint4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(uint4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(uint4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(uint4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(uint4);
-ushort4 __ovld __cnfn convert_ushort4(uint4);
-ushort4 __ovld __cnfn convert_ushort4_sat(uint4);
-ushort4 __ovld __cnfn convert_ushort4_rte(long4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(long4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(long4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(long4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(long4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(long4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(long4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(long4);
-ushort4 __ovld __cnfn convert_ushort4(long4);
-ushort4 __ovld __cnfn convert_ushort4_sat(long4);
-ushort4 __ovld __cnfn convert_ushort4_rte(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(ulong4);
-ushort4 __ovld __cnfn convert_ushort4(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_sat(ulong4);
-ushort4 __ovld __cnfn convert_ushort4_rte(float4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(float4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(float4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(float4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(float4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(float4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(float4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(float4);
-ushort4 __ovld __cnfn convert_ushort4(float4);
-ushort4 __ovld __cnfn convert_ushort4_sat(float4);
-int4 __ovld __cnfn convert_int4_rte(char4);
-int4 __ovld __cnfn convert_int4_sat_rte(char4);
-int4 __ovld __cnfn convert_int4_rtz(char4);
-int4 __ovld __cnfn convert_int4_sat_rtz(char4);
-int4 __ovld __cnfn convert_int4_rtp(char4);
-int4 __ovld __cnfn convert_int4_sat_rtp(char4);
-int4 __ovld __cnfn convert_int4_rtn(char4);
-int4 __ovld __cnfn convert_int4_sat_rtn(char4);
-int4 __ovld __cnfn convert_int4(char4);
-int4 __ovld __cnfn convert_int4_sat(char4);
-int4 __ovld __cnfn convert_int4_rte(uchar4);
-int4 __ovld __cnfn convert_int4_sat_rte(uchar4);
-int4 __ovld __cnfn convert_int4_rtz(uchar4);
-int4 __ovld __cnfn convert_int4_sat_rtz(uchar4);
-int4 __ovld __cnfn convert_int4_rtp(uchar4);
-int4 __ovld __cnfn convert_int4_sat_rtp(uchar4);
-int4 __ovld __cnfn convert_int4_rtn(uchar4);
-int4 __ovld __cnfn convert_int4_sat_rtn(uchar4);
-int4 __ovld __cnfn convert_int4(uchar4);
-int4 __ovld __cnfn convert_int4_sat(uchar4);
-int4 __ovld __cnfn convert_int4_rte(short4);
-int4 __ovld __cnfn convert_int4_sat_rte(short4);
-int4 __ovld __cnfn convert_int4_rtz(short4);
-int4 __ovld __cnfn convert_int4_sat_rtz(short4);
-int4 __ovld __cnfn convert_int4_rtp(short4);
-int4 __ovld __cnfn convert_int4_sat_rtp(short4);
-int4 __ovld __cnfn convert_int4_rtn(short4);
-int4 __ovld __cnfn convert_int4_sat_rtn(short4);
-int4 __ovld __cnfn convert_int4(short4);
-int4 __ovld __cnfn convert_int4_sat(short4);
-int4 __ovld __cnfn convert_int4_rte(ushort4);
-int4 __ovld __cnfn convert_int4_sat_rte(ushort4);
-int4 __ovld __cnfn convert_int4_rtz(ushort4);
-int4 __ovld __cnfn convert_int4_sat_rtz(ushort4);
-int4 __ovld __cnfn convert_int4_rtp(ushort4);
-int4 __ovld __cnfn convert_int4_sat_rtp(ushort4);
-int4 __ovld __cnfn convert_int4_rtn(ushort4);
-int4 __ovld __cnfn convert_int4_sat_rtn(ushort4);
-int4 __ovld __cnfn convert_int4(ushort4);
-int4 __ovld __cnfn convert_int4_sat(ushort4);
-int4 __ovld __cnfn convert_int4_rte(int4);
-int4 __ovld __cnfn convert_int4_sat_rte(int4);
-int4 __ovld __cnfn convert_int4_rtz(int4);
-int4 __ovld __cnfn convert_int4_sat_rtz(int4);
-int4 __ovld __cnfn convert_int4_rtp(int4);
-int4 __ovld __cnfn convert_int4_sat_rtp(int4);
-int4 __ovld __cnfn convert_int4_rtn(int4);
-int4 __ovld __cnfn convert_int4_sat_rtn(int4);
-int4 __ovld __cnfn convert_int4(int4);
-int4 __ovld __cnfn convert_int4_sat(int4);
-int4 __ovld __cnfn convert_int4_rte(uint4);
-int4 __ovld __cnfn convert_int4_sat_rte(uint4);
-int4 __ovld __cnfn convert_int4_rtz(uint4);
-int4 __ovld __cnfn convert_int4_sat_rtz(uint4);
-int4 __ovld __cnfn convert_int4_rtp(uint4);
-int4 __ovld __cnfn convert_int4_sat_rtp(uint4);
-int4 __ovld __cnfn convert_int4_rtn(uint4);
-int4 __ovld __cnfn convert_int4_sat_rtn(uint4);
-int4 __ovld __cnfn convert_int4(uint4);
-int4 __ovld __cnfn convert_int4_sat(uint4);
-int4 __ovld __cnfn convert_int4_rte(long4);
-int4 __ovld __cnfn convert_int4_sat_rte(long4);
-int4 __ovld __cnfn convert_int4_rtz(long4);
-int4 __ovld __cnfn convert_int4_sat_rtz(long4);
-int4 __ovld __cnfn convert_int4_rtp(long4);
-int4 __ovld __cnfn convert_int4_sat_rtp(long4);
-int4 __ovld __cnfn convert_int4_rtn(long4);
-int4 __ovld __cnfn convert_int4_sat_rtn(long4);
-int4 __ovld __cnfn convert_int4(long4);
-int4 __ovld __cnfn convert_int4_sat(long4);
-int4 __ovld __cnfn convert_int4_rte(ulong4);
-int4 __ovld __cnfn convert_int4_sat_rte(ulong4);
-int4 __ovld __cnfn convert_int4_rtz(ulong4);
-int4 __ovld __cnfn convert_int4_sat_rtz(ulong4);
-int4 __ovld __cnfn convert_int4_rtp(ulong4);
-int4 __ovld __cnfn convert_int4_sat_rtp(ulong4);
-int4 __ovld __cnfn convert_int4_rtn(ulong4);
-int4 __ovld __cnfn convert_int4_sat_rtn(ulong4);
-int4 __ovld __cnfn convert_int4(ulong4);
-int4 __ovld __cnfn convert_int4_sat(ulong4);
-int4 __ovld __cnfn convert_int4_rte(float4);
-int4 __ovld __cnfn convert_int4_sat_rte(float4);
-int4 __ovld __cnfn convert_int4_rtz(float4);
-int4 __ovld __cnfn convert_int4_sat_rtz(float4);
-int4 __ovld __cnfn convert_int4_rtp(float4);
-int4 __ovld __cnfn convert_int4_sat_rtp(float4);
-int4 __ovld __cnfn convert_int4_rtn(float4);
-int4 __ovld __cnfn convert_int4_sat_rtn(float4);
-int4 __ovld __cnfn convert_int4(float4);
-int4 __ovld __cnfn convert_int4_sat(float4);
-uint4 __ovld __cnfn convert_uint4_rte(char4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(char4);
-uint4 __ovld __cnfn convert_uint4_rtz(char4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(char4);
-uint4 __ovld __cnfn convert_uint4_rtp(char4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(char4);
-uint4 __ovld __cnfn convert_uint4_rtn(char4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(char4);
-uint4 __ovld __cnfn convert_uint4(char4);
-uint4 __ovld __cnfn convert_uint4_sat(char4);
-uint4 __ovld __cnfn convert_uint4_rte(uchar4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(uchar4);
-uint4 __ovld __cnfn convert_uint4_rtz(uchar4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(uchar4);
-uint4 __ovld __cnfn convert_uint4_rtp(uchar4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(uchar4);
-uint4 __ovld __cnfn convert_uint4_rtn(uchar4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(uchar4);
-uint4 __ovld __cnfn convert_uint4(uchar4);
-uint4 __ovld __cnfn convert_uint4_sat(uchar4);
-uint4 __ovld __cnfn convert_uint4_rte(short4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(short4);
-uint4 __ovld __cnfn convert_uint4_rtz(short4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(short4);
-uint4 __ovld __cnfn convert_uint4_rtp(short4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(short4);
-uint4 __ovld __cnfn convert_uint4_rtn(short4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(short4);
-uint4 __ovld __cnfn convert_uint4(short4);
-uint4 __ovld __cnfn convert_uint4_sat(short4);
-uint4 __ovld __cnfn convert_uint4_rte(ushort4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(ushort4);
-uint4 __ovld __cnfn convert_uint4_rtz(ushort4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(ushort4);
-uint4 __ovld __cnfn convert_uint4_rtp(ushort4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(ushort4);
-uint4 __ovld __cnfn convert_uint4_rtn(ushort4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(ushort4);
-uint4 __ovld __cnfn convert_uint4(ushort4);
-uint4 __ovld __cnfn convert_uint4_sat(ushort4);
-uint4 __ovld __cnfn convert_uint4_rte(int4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(int4);
-uint4 __ovld __cnfn convert_uint4_rtz(int4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(int4);
-uint4 __ovld __cnfn convert_uint4_rtp(int4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(int4);
-uint4 __ovld __cnfn convert_uint4_rtn(int4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(int4);
-uint4 __ovld __cnfn convert_uint4(int4);
-uint4 __ovld __cnfn convert_uint4_sat(int4);
-uint4 __ovld __cnfn convert_uint4_rte(uint4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(uint4);
-uint4 __ovld __cnfn convert_uint4_rtz(uint4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(uint4);
-uint4 __ovld __cnfn convert_uint4_rtp(uint4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(uint4);
-uint4 __ovld __cnfn convert_uint4_rtn(uint4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(uint4);
-uint4 __ovld __cnfn convert_uint4(uint4);
-uint4 __ovld __cnfn convert_uint4_sat(uint4);
-uint4 __ovld __cnfn convert_uint4_rte(long4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(long4);
-uint4 __ovld __cnfn convert_uint4_rtz(long4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(long4);
-uint4 __ovld __cnfn convert_uint4_rtp(long4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(long4);
-uint4 __ovld __cnfn convert_uint4_rtn(long4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(long4);
-uint4 __ovld __cnfn convert_uint4(long4);
-uint4 __ovld __cnfn convert_uint4_sat(long4);
-uint4 __ovld __cnfn convert_uint4_rte(ulong4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(ulong4);
-uint4 __ovld __cnfn convert_uint4_rtz(ulong4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(ulong4);
-uint4 __ovld __cnfn convert_uint4_rtp(ulong4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(ulong4);
-uint4 __ovld __cnfn convert_uint4_rtn(ulong4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(ulong4);
-uint4 __ovld __cnfn convert_uint4(ulong4);
-uint4 __ovld __cnfn convert_uint4_sat(ulong4);
-uint4 __ovld __cnfn convert_uint4_rte(float4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(float4);
-uint4 __ovld __cnfn convert_uint4_rtz(float4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(float4);
-uint4 __ovld __cnfn convert_uint4_rtp(float4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(float4);
-uint4 __ovld __cnfn convert_uint4_rtn(float4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(float4);
-uint4 __ovld __cnfn convert_uint4(float4);
-uint4 __ovld __cnfn convert_uint4_sat(float4);
-long4 __ovld __cnfn convert_long4_rte(char4);
-long4 __ovld __cnfn convert_long4_sat_rte(char4);
-long4 __ovld __cnfn convert_long4_rtz(char4);
-long4 __ovld __cnfn convert_long4_sat_rtz(char4);
-long4 __ovld __cnfn convert_long4_rtp(char4);
-long4 __ovld __cnfn convert_long4_sat_rtp(char4);
-long4 __ovld __cnfn convert_long4_rtn(char4);
-long4 __ovld __cnfn convert_long4_sat_rtn(char4);
-long4 __ovld __cnfn convert_long4(char4);
-long4 __ovld __cnfn convert_long4_sat(char4);
-long4 __ovld __cnfn convert_long4_rte(uchar4);
-long4 __ovld __cnfn convert_long4_sat_rte(uchar4);
-long4 __ovld __cnfn convert_long4_rtz(uchar4);
-long4 __ovld __cnfn convert_long4_sat_rtz(uchar4);
-long4 __ovld __cnfn convert_long4_rtp(uchar4);
-long4 __ovld __cnfn convert_long4_sat_rtp(uchar4);
-long4 __ovld __cnfn convert_long4_rtn(uchar4);
-long4 __ovld __cnfn convert_long4_sat_rtn(uchar4);
-long4 __ovld __cnfn convert_long4(uchar4);
-long4 __ovld __cnfn convert_long4_sat(uchar4);
-long4 __ovld __cnfn convert_long4_rte(short4);
-long4 __ovld __cnfn convert_long4_sat_rte(short4);
-long4 __ovld __cnfn convert_long4_rtz(short4);
-long4 __ovld __cnfn convert_long4_sat_rtz(short4);
-long4 __ovld __cnfn convert_long4_rtp(short4);
-long4 __ovld __cnfn convert_long4_sat_rtp(short4);
-long4 __ovld __cnfn convert_long4_rtn(short4);
-long4 __ovld __cnfn convert_long4_sat_rtn(short4);
-long4 __ovld __cnfn convert_long4(short4);
-long4 __ovld __cnfn convert_long4_sat(short4);
-long4 __ovld __cnfn convert_long4_rte(ushort4);
-long4 __ovld __cnfn convert_long4_sat_rte(ushort4);
-long4 __ovld __cnfn convert_long4_rtz(ushort4);
-long4 __ovld __cnfn convert_long4_sat_rtz(ushort4);
-long4 __ovld __cnfn convert_long4_rtp(ushort4);
-long4 __ovld __cnfn convert_long4_sat_rtp(ushort4);
-long4 __ovld __cnfn convert_long4_rtn(ushort4);
-long4 __ovld __cnfn convert_long4_sat_rtn(ushort4);
-long4 __ovld __cnfn convert_long4(ushort4);
-long4 __ovld __cnfn convert_long4_sat(ushort4);
-long4 __ovld __cnfn convert_long4_rte(int4);
-long4 __ovld __cnfn convert_long4_sat_rte(int4);
-long4 __ovld __cnfn convert_long4_rtz(int4);
-long4 __ovld __cnfn convert_long4_sat_rtz(int4);
-long4 __ovld __cnfn convert_long4_rtp(int4);
-long4 __ovld __cnfn convert_long4_sat_rtp(int4);
-long4 __ovld __cnfn convert_long4_rtn(int4);
-long4 __ovld __cnfn convert_long4_sat_rtn(int4);
-long4 __ovld __cnfn convert_long4(int4);
-long4 __ovld __cnfn convert_long4_sat(int4);
-long4 __ovld __cnfn convert_long4_rte(uint4);
-long4 __ovld __cnfn convert_long4_sat_rte(uint4);
-long4 __ovld __cnfn convert_long4_rtz(uint4);
-long4 __ovld __cnfn convert_long4_sat_rtz(uint4);
-long4 __ovld __cnfn convert_long4_rtp(uint4);
-long4 __ovld __cnfn convert_long4_sat_rtp(uint4);
-long4 __ovld __cnfn convert_long4_rtn(uint4);
-long4 __ovld __cnfn convert_long4_sat_rtn(uint4);
-long4 __ovld __cnfn convert_long4(uint4);
-long4 __ovld __cnfn convert_long4_sat(uint4);
-long4 __ovld __cnfn convert_long4_rte(long4);
-long4 __ovld __cnfn convert_long4_sat_rte(long4);
-long4 __ovld __cnfn convert_long4_rtz(long4);
-long4 __ovld __cnfn convert_long4_sat_rtz(long4);
-long4 __ovld __cnfn convert_long4_rtp(long4);
-long4 __ovld __cnfn convert_long4_sat_rtp(long4);
-long4 __ovld __cnfn convert_long4_rtn(long4);
-long4 __ovld __cnfn convert_long4_sat_rtn(long4);
-long4 __ovld __cnfn convert_long4(long4);
-long4 __ovld __cnfn convert_long4_sat(long4);
-long4 __ovld __cnfn convert_long4_rte(ulong4);
-long4 __ovld __cnfn convert_long4_sat_rte(ulong4);
-long4 __ovld __cnfn convert_long4_rtz(ulong4);
-long4 __ovld __cnfn convert_long4_sat_rtz(ulong4);
-long4 __ovld __cnfn convert_long4_rtp(ulong4);
-long4 __ovld __cnfn convert_long4_sat_rtp(ulong4);
-long4 __ovld __cnfn convert_long4_rtn(ulong4);
-long4 __ovld __cnfn convert_long4_sat_rtn(ulong4);
-long4 __ovld __cnfn convert_long4(ulong4);
-long4 __ovld __cnfn convert_long4_sat(ulong4);
-long4 __ovld __cnfn convert_long4_rte(float4);
-long4 __ovld __cnfn convert_long4_sat_rte(float4);
-long4 __ovld __cnfn convert_long4_rtz(float4);
-long4 __ovld __cnfn convert_long4_sat_rtz(float4);
-long4 __ovld __cnfn convert_long4_rtp(float4);
-long4 __ovld __cnfn convert_long4_sat_rtp(float4);
-long4 __ovld __cnfn convert_long4_rtn(float4);
-long4 __ovld __cnfn convert_long4_sat_rtn(float4);
-long4 __ovld __cnfn convert_long4(float4);
-long4 __ovld __cnfn convert_long4_sat(float4);
-ulong4 __ovld __cnfn convert_ulong4_rte(char4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(char4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(char4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(char4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(char4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(char4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(char4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(char4);
-ulong4 __ovld __cnfn convert_ulong4(char4);
-ulong4 __ovld __cnfn convert_ulong4_sat(char4);
-ulong4 __ovld __cnfn convert_ulong4_rte(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(uchar4);
-ulong4 __ovld __cnfn convert_ulong4(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_sat(uchar4);
-ulong4 __ovld __cnfn convert_ulong4_rte(short4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(short4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(short4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(short4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(short4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(short4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(short4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(short4);
-ulong4 __ovld __cnfn convert_ulong4(short4);
-ulong4 __ovld __cnfn convert_ulong4_sat(short4);
-ulong4 __ovld __cnfn convert_ulong4_rte(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(ushort4);
-ulong4 __ovld __cnfn convert_ulong4(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_sat(ushort4);
-ulong4 __ovld __cnfn convert_ulong4_rte(int4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(int4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(int4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(int4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(int4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(int4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(int4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(int4);
-ulong4 __ovld __cnfn convert_ulong4(int4);
-ulong4 __ovld __cnfn convert_ulong4_sat(int4);
-ulong4 __ovld __cnfn convert_ulong4_rte(uint4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(uint4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(uint4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(uint4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(uint4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(uint4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(uint4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(uint4);
-ulong4 __ovld __cnfn convert_ulong4(uint4);
-ulong4 __ovld __cnfn convert_ulong4_sat(uint4);
-ulong4 __ovld __cnfn convert_ulong4_rte(long4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(long4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(long4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(long4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(long4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(long4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(long4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(long4);
-ulong4 __ovld __cnfn convert_ulong4(long4);
-ulong4 __ovld __cnfn convert_ulong4_sat(long4);
-ulong4 __ovld __cnfn convert_ulong4_rte(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(ulong4);
-ulong4 __ovld __cnfn convert_ulong4(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_sat(ulong4);
-ulong4 __ovld __cnfn convert_ulong4_rte(float4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(float4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(float4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(float4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(float4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(float4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(float4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(float4);
-ulong4 __ovld __cnfn convert_ulong4(float4);
-ulong4 __ovld __cnfn convert_ulong4_sat(float4);
-float4 __ovld __cnfn convert_float4_rte(char4);
-float4 __ovld __cnfn convert_float4_rtz(char4);
-float4 __ovld __cnfn convert_float4_rtp(char4);
-float4 __ovld __cnfn convert_float4_rtn(char4);
-float4 __ovld __cnfn convert_float4(char4);
-float4 __ovld __cnfn convert_float4_rte(uchar4);
-float4 __ovld __cnfn convert_float4_rtz(uchar4);
-float4 __ovld __cnfn convert_float4_rtp(uchar4);
-float4 __ovld __cnfn convert_float4_rtn(uchar4);
-float4 __ovld __cnfn convert_float4(uchar4);
-float4 __ovld __cnfn convert_float4_rte(short4);
-float4 __ovld __cnfn convert_float4_rtz(short4);
-float4 __ovld __cnfn convert_float4_rtp(short4);
-float4 __ovld __cnfn convert_float4_rtn(short4);
-float4 __ovld __cnfn convert_float4(short4);
-float4 __ovld __cnfn convert_float4_rte(ushort4);
-float4 __ovld __cnfn convert_float4_rtz(ushort4);
-float4 __ovld __cnfn convert_float4_rtp(ushort4);
-float4 __ovld __cnfn convert_float4_rtn(ushort4);
-float4 __ovld __cnfn convert_float4(ushort4);
-float4 __ovld __cnfn convert_float4_rte(int4);
-float4 __ovld __cnfn convert_float4_rtz(int4);
-float4 __ovld __cnfn convert_float4_rtp(int4);
-float4 __ovld __cnfn convert_float4_rtn(int4);
-float4 __ovld __cnfn convert_float4(int4);
-float4 __ovld __cnfn convert_float4_rte(uint4);
-float4 __ovld __cnfn convert_float4_rtz(uint4);
-float4 __ovld __cnfn convert_float4_rtp(uint4);
-float4 __ovld __cnfn convert_float4_rtn(uint4);
-float4 __ovld __cnfn convert_float4(uint4);
-float4 __ovld __cnfn convert_float4_rte(long4);
-float4 __ovld __cnfn convert_float4_rtz(long4);
-float4 __ovld __cnfn convert_float4_rtp(long4);
-float4 __ovld __cnfn convert_float4_rtn(long4);
-float4 __ovld __cnfn convert_float4(long4);
-float4 __ovld __cnfn convert_float4_rte(ulong4);
-float4 __ovld __cnfn convert_float4_rtz(ulong4);
-float4 __ovld __cnfn convert_float4_rtp(ulong4);
-float4 __ovld __cnfn convert_float4_rtn(ulong4);
-float4 __ovld __cnfn convert_float4(ulong4);
-float4 __ovld __cnfn convert_float4_rte(float4);
-float4 __ovld __cnfn convert_float4_rtz(float4);
-float4 __ovld __cnfn convert_float4_rtp(float4);
-float4 __ovld __cnfn convert_float4_rtn(float4);
-float4 __ovld __cnfn convert_float4(float4);
-char8 __ovld __cnfn convert_char8_rte(char8);
-char8 __ovld __cnfn convert_char8_sat_rte(char8);
-char8 __ovld __cnfn convert_char8_rtz(char8);
-char8 __ovld __cnfn convert_char8_sat_rtz(char8);
-char8 __ovld __cnfn convert_char8_rtp(char8);
-char8 __ovld __cnfn convert_char8_sat_rtp(char8);
-char8 __ovld __cnfn convert_char8_rtn(char8);
-char8 __ovld __cnfn convert_char8_sat_rtn(char8);
-char8 __ovld __cnfn convert_char8(char8);
-char8 __ovld __cnfn convert_char8_sat(char8);
-char8 __ovld __cnfn convert_char8_rte(uchar8);
-char8 __ovld __cnfn convert_char8_sat_rte(uchar8);
-char8 __ovld __cnfn convert_char8_rtz(uchar8);
-char8 __ovld __cnfn convert_char8_sat_rtz(uchar8);
-char8 __ovld __cnfn convert_char8_rtp(uchar8);
-char8 __ovld __cnfn convert_char8_sat_rtp(uchar8);
-char8 __ovld __cnfn convert_char8_rtn(uchar8);
-char8 __ovld __cnfn convert_char8_sat_rtn(uchar8);
-char8 __ovld __cnfn convert_char8(uchar8);
-char8 __ovld __cnfn convert_char8_sat(uchar8);
-char8 __ovld __cnfn convert_char8_rte(short8);
-char8 __ovld __cnfn convert_char8_sat_rte(short8);
-char8 __ovld __cnfn convert_char8_rtz(short8);
-char8 __ovld __cnfn convert_char8_sat_rtz(short8);
-char8 __ovld __cnfn convert_char8_rtp(short8);
-char8 __ovld __cnfn convert_char8_sat_rtp(short8);
-char8 __ovld __cnfn convert_char8_rtn(short8);
-char8 __ovld __cnfn convert_char8_sat_rtn(short8);
-char8 __ovld __cnfn convert_char8(short8);
-char8 __ovld __cnfn convert_char8_sat(short8);
-char8 __ovld __cnfn convert_char8_rte(ushort8);
-char8 __ovld __cnfn convert_char8_sat_rte(ushort8);
-char8 __ovld __cnfn convert_char8_rtz(ushort8);
-char8 __ovld __cnfn convert_char8_sat_rtz(ushort8);
-char8 __ovld __cnfn convert_char8_rtp(ushort8);
-char8 __ovld __cnfn convert_char8_sat_rtp(ushort8);
-char8 __ovld __cnfn convert_char8_rtn(ushort8);
-char8 __ovld __cnfn convert_char8_sat_rtn(ushort8);
-char8 __ovld __cnfn convert_char8(ushort8);
-char8 __ovld __cnfn convert_char8_sat(ushort8);
-char8 __ovld __cnfn convert_char8_rte(int8);
-char8 __ovld __cnfn convert_char8_sat_rte(int8);
-char8 __ovld __cnfn convert_char8_rtz(int8);
-char8 __ovld __cnfn convert_char8_sat_rtz(int8);
-char8 __ovld __cnfn convert_char8_rtp(int8);
-char8 __ovld __cnfn convert_char8_sat_rtp(int8);
-char8 __ovld __cnfn convert_char8_rtn(int8);
-char8 __ovld __cnfn convert_char8_sat_rtn(int8);
-char8 __ovld __cnfn convert_char8(int8);
-char8 __ovld __cnfn convert_char8_sat(int8);
-char8 __ovld __cnfn convert_char8_rte(uint8);
-char8 __ovld __cnfn convert_char8_sat_rte(uint8);
-char8 __ovld __cnfn convert_char8_rtz(uint8);
-char8 __ovld __cnfn convert_char8_sat_rtz(uint8);
-char8 __ovld __cnfn convert_char8_rtp(uint8);
-char8 __ovld __cnfn convert_char8_sat_rtp(uint8);
-char8 __ovld __cnfn convert_char8_rtn(uint8);
-char8 __ovld __cnfn convert_char8_sat_rtn(uint8);
-char8 __ovld __cnfn convert_char8(uint8);
-char8 __ovld __cnfn convert_char8_sat(uint8);
-char8 __ovld __cnfn convert_char8_rte(long8);
-char8 __ovld __cnfn convert_char8_sat_rte(long8);
-char8 __ovld __cnfn convert_char8_rtz(long8);
-char8 __ovld __cnfn convert_char8_sat_rtz(long8);
-char8 __ovld __cnfn convert_char8_rtp(long8);
-char8 __ovld __cnfn convert_char8_sat_rtp(long8);
-char8 __ovld __cnfn convert_char8_rtn(long8);
-char8 __ovld __cnfn convert_char8_sat_rtn(long8);
-char8 __ovld __cnfn convert_char8(long8);
-char8 __ovld __cnfn convert_char8_sat(long8);
-char8 __ovld __cnfn convert_char8_rte(ulong8);
-char8 __ovld __cnfn convert_char8_sat_rte(ulong8);
-char8 __ovld __cnfn convert_char8_rtz(ulong8);
-char8 __ovld __cnfn convert_char8_sat_rtz(ulong8);
-char8 __ovld __cnfn convert_char8_rtp(ulong8);
-char8 __ovld __cnfn convert_char8_sat_rtp(ulong8);
-char8 __ovld __cnfn convert_char8_rtn(ulong8);
-char8 __ovld __cnfn convert_char8_sat_rtn(ulong8);
-char8 __ovld __cnfn convert_char8(ulong8);
-char8 __ovld __cnfn convert_char8_sat(ulong8);
-char8 __ovld __cnfn convert_char8_rte(float8);
-char8 __ovld __cnfn convert_char8_sat_rte(float8);
-char8 __ovld __cnfn convert_char8_rtz(float8);
-char8 __ovld __cnfn convert_char8_sat_rtz(float8);
-char8 __ovld __cnfn convert_char8_rtp(float8);
-char8 __ovld __cnfn convert_char8_sat_rtp(float8);
-char8 __ovld __cnfn convert_char8_rtn(float8);
-char8 __ovld __cnfn convert_char8_sat_rtn(float8);
-char8 __ovld __cnfn convert_char8(float8);
-char8 __ovld __cnfn convert_char8_sat(float8);
-uchar8 __ovld __cnfn convert_uchar8_rte(char8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(char8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(char8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(char8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(char8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(char8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(char8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(char8);
-uchar8 __ovld __cnfn convert_uchar8(char8);
-uchar8 __ovld __cnfn convert_uchar8_sat(char8);
-uchar8 __ovld __cnfn convert_uchar8_rte(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(uchar8);
-uchar8 __ovld __cnfn convert_uchar8(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_sat(uchar8);
-uchar8 __ovld __cnfn convert_uchar8_rte(short8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(short8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(short8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(short8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(short8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(short8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(short8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(short8);
-uchar8 __ovld __cnfn convert_uchar8(short8);
-uchar8 __ovld __cnfn convert_uchar8_sat(short8);
-uchar8 __ovld __cnfn convert_uchar8_rte(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(ushort8);
-uchar8 __ovld __cnfn convert_uchar8(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_sat(ushort8);
-uchar8 __ovld __cnfn convert_uchar8_rte(int8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(int8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(int8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(int8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(int8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(int8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(int8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(int8);
-uchar8 __ovld __cnfn convert_uchar8(int8);
-uchar8 __ovld __cnfn convert_uchar8_sat(int8);
-uchar8 __ovld __cnfn convert_uchar8_rte(uint8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(uint8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(uint8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(uint8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(uint8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(uint8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(uint8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(uint8);
-uchar8 __ovld __cnfn convert_uchar8(uint8);
-uchar8 __ovld __cnfn convert_uchar8_sat(uint8);
-uchar8 __ovld __cnfn convert_uchar8_rte(long8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(long8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(long8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(long8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(long8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(long8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(long8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(long8);
-uchar8 __ovld __cnfn convert_uchar8(long8);
-uchar8 __ovld __cnfn convert_uchar8_sat(long8);
-uchar8 __ovld __cnfn convert_uchar8_rte(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(ulong8);
-uchar8 __ovld __cnfn convert_uchar8(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_sat(ulong8);
-uchar8 __ovld __cnfn convert_uchar8_rte(float8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(float8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(float8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(float8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(float8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(float8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(float8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(float8);
-uchar8 __ovld __cnfn convert_uchar8(float8);
-uchar8 __ovld __cnfn convert_uchar8_sat(float8);
-short8 __ovld __cnfn convert_short8_rte(char8);
-short8 __ovld __cnfn convert_short8_sat_rte(char8);
-short8 __ovld __cnfn convert_short8_rtz(char8);
-short8 __ovld __cnfn convert_short8_sat_rtz(char8);
-short8 __ovld __cnfn convert_short8_rtp(char8);
-short8 __ovld __cnfn convert_short8_sat_rtp(char8);
-short8 __ovld __cnfn convert_short8_rtn(char8);
-short8 __ovld __cnfn convert_short8_sat_rtn(char8);
-short8 __ovld __cnfn convert_short8(char8);
-short8 __ovld __cnfn convert_short8_sat(char8);
-short8 __ovld __cnfn convert_short8_rte(uchar8);
-short8 __ovld __cnfn convert_short8_sat_rte(uchar8);
-short8 __ovld __cnfn convert_short8_rtz(uchar8);
-short8 __ovld __cnfn convert_short8_sat_rtz(uchar8);
-short8 __ovld __cnfn convert_short8_rtp(uchar8);
-short8 __ovld __cnfn convert_short8_sat_rtp(uchar8);
-short8 __ovld __cnfn convert_short8_rtn(uchar8);
-short8 __ovld __cnfn convert_short8_sat_rtn(uchar8);
-short8 __ovld __cnfn convert_short8(uchar8);
-short8 __ovld __cnfn convert_short8_sat(uchar8);
-short8 __ovld __cnfn convert_short8_rte(short8);
-short8 __ovld __cnfn convert_short8_sat_rte(short8);
-short8 __ovld __cnfn convert_short8_rtz(short8);
-short8 __ovld __cnfn convert_short8_sat_rtz(short8);
-short8 __ovld __cnfn convert_short8_rtp(short8);
-short8 __ovld __cnfn convert_short8_sat_rtp(short8);
-short8 __ovld __cnfn convert_short8_rtn(short8);
-short8 __ovld __cnfn convert_short8_sat_rtn(short8);
-short8 __ovld __cnfn convert_short8(short8);
-short8 __ovld __cnfn convert_short8_sat(short8);
-short8 __ovld __cnfn convert_short8_rte(ushort8);
-short8 __ovld __cnfn convert_short8_sat_rte(ushort8);
-short8 __ovld __cnfn convert_short8_rtz(ushort8);
-short8 __ovld __cnfn convert_short8_sat_rtz(ushort8);
-short8 __ovld __cnfn convert_short8_rtp(ushort8);
-short8 __ovld __cnfn convert_short8_sat_rtp(ushort8);
-short8 __ovld __cnfn convert_short8_rtn(ushort8);
-short8 __ovld __cnfn convert_short8_sat_rtn(ushort8);
-short8 __ovld __cnfn convert_short8(ushort8);
-short8 __ovld __cnfn convert_short8_sat(ushort8);
-short8 __ovld __cnfn convert_short8_rte(int8);
-short8 __ovld __cnfn convert_short8_sat_rte(int8);
-short8 __ovld __cnfn convert_short8_rtz(int8);
-short8 __ovld __cnfn convert_short8_sat_rtz(int8);
-short8 __ovld __cnfn convert_short8_rtp(int8);
-short8 __ovld __cnfn convert_short8_sat_rtp(int8);
-short8 __ovld __cnfn convert_short8_rtn(int8);
-short8 __ovld __cnfn convert_short8_sat_rtn(int8);
-short8 __ovld __cnfn convert_short8(int8);
-short8 __ovld __cnfn convert_short8_sat(int8);
-short8 __ovld __cnfn convert_short8_rte(uint8);
-short8 __ovld __cnfn convert_short8_sat_rte(uint8);
-short8 __ovld __cnfn convert_short8_rtz(uint8);
-short8 __ovld __cnfn convert_short8_sat_rtz(uint8);
-short8 __ovld __cnfn convert_short8_rtp(uint8);
-short8 __ovld __cnfn convert_short8_sat_rtp(uint8);
-short8 __ovld __cnfn convert_short8_rtn(uint8);
-short8 __ovld __cnfn convert_short8_sat_rtn(uint8);
-short8 __ovld __cnfn convert_short8(uint8);
-short8 __ovld __cnfn convert_short8_sat(uint8);
-short8 __ovld __cnfn convert_short8_rte(long8);
-short8 __ovld __cnfn convert_short8_sat_rte(long8);
-short8 __ovld __cnfn convert_short8_rtz(long8);
-short8 __ovld __cnfn convert_short8_sat_rtz(long8);
-short8 __ovld __cnfn convert_short8_rtp(long8);
-short8 __ovld __cnfn convert_short8_sat_rtp(long8);
-short8 __ovld __cnfn convert_short8_rtn(long8);
-short8 __ovld __cnfn convert_short8_sat_rtn(long8);
-short8 __ovld __cnfn convert_short8(long8);
-short8 __ovld __cnfn convert_short8_sat(long8);
-short8 __ovld __cnfn convert_short8_rte(ulong8);
-short8 __ovld __cnfn convert_short8_sat_rte(ulong8);
-short8 __ovld __cnfn convert_short8_rtz(ulong8);
-short8 __ovld __cnfn convert_short8_sat_rtz(ulong8);
-short8 __ovld __cnfn convert_short8_rtp(ulong8);
-short8 __ovld __cnfn convert_short8_sat_rtp(ulong8);
-short8 __ovld __cnfn convert_short8_rtn(ulong8);
-short8 __ovld __cnfn convert_short8_sat_rtn(ulong8);
-short8 __ovld __cnfn convert_short8(ulong8);
-short8 __ovld __cnfn convert_short8_sat(ulong8);
-short8 __ovld __cnfn convert_short8_rte(float8);
-short8 __ovld __cnfn convert_short8_sat_rte(float8);
-short8 __ovld __cnfn convert_short8_rtz(float8);
-short8 __ovld __cnfn convert_short8_sat_rtz(float8);
-short8 __ovld __cnfn convert_short8_rtp(float8);
-short8 __ovld __cnfn convert_short8_sat_rtp(float8);
-short8 __ovld __cnfn convert_short8_rtn(float8);
-short8 __ovld __cnfn convert_short8_sat_rtn(float8);
-short8 __ovld __cnfn convert_short8(float8);
-short8 __ovld __cnfn convert_short8_sat(float8);
-ushort8 __ovld __cnfn convert_ushort8_rte(char8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(char8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(char8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(char8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(char8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(char8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(char8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(char8);
-ushort8 __ovld __cnfn convert_ushort8(char8);
-ushort8 __ovld __cnfn convert_ushort8_sat(char8);
-ushort8 __ovld __cnfn convert_ushort8_rte(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(uchar8);
-ushort8 __ovld __cnfn convert_ushort8(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_sat(uchar8);
-ushort8 __ovld __cnfn convert_ushort8_rte(short8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(short8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(short8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(short8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(short8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(short8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(short8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(short8);
-ushort8 __ovld __cnfn convert_ushort8(short8);
-ushort8 __ovld __cnfn convert_ushort8_sat(short8);
-ushort8 __ovld __cnfn convert_ushort8_rte(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(ushort8);
-ushort8 __ovld __cnfn convert_ushort8(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_sat(ushort8);
-ushort8 __ovld __cnfn convert_ushort8_rte(int8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(int8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(int8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(int8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(int8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(int8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(int8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(int8);
-ushort8 __ovld __cnfn convert_ushort8(int8);
-ushort8 __ovld __cnfn convert_ushort8_sat(int8);
-ushort8 __ovld __cnfn convert_ushort8_rte(uint8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(uint8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(uint8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(uint8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(uint8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(uint8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(uint8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(uint8);
-ushort8 __ovld __cnfn convert_ushort8(uint8);
-ushort8 __ovld __cnfn convert_ushort8_sat(uint8);
-ushort8 __ovld __cnfn convert_ushort8_rte(long8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(long8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(long8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(long8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(long8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(long8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(long8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(long8);
-ushort8 __ovld __cnfn convert_ushort8(long8);
-ushort8 __ovld __cnfn convert_ushort8_sat(long8);
-ushort8 __ovld __cnfn convert_ushort8_rte(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(ulong8);
-ushort8 __ovld __cnfn convert_ushort8(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_sat(ulong8);
-ushort8 __ovld __cnfn convert_ushort8_rte(float8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(float8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(float8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(float8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(float8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(float8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(float8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(float8);
-ushort8 __ovld __cnfn convert_ushort8(float8);
-ushort8 __ovld __cnfn convert_ushort8_sat(float8);
-int8 __ovld __cnfn convert_int8_rte(char8);
-int8 __ovld __cnfn convert_int8_sat_rte(char8);
-int8 __ovld __cnfn convert_int8_rtz(char8);
-int8 __ovld __cnfn convert_int8_sat_rtz(char8);
-int8 __ovld __cnfn convert_int8_rtp(char8);
-int8 __ovld __cnfn convert_int8_sat_rtp(char8);
-int8 __ovld __cnfn convert_int8_rtn(char8);
-int8 __ovld __cnfn convert_int8_sat_rtn(char8);
-int8 __ovld __cnfn convert_int8(char8);
-int8 __ovld __cnfn convert_int8_sat(char8);
-int8 __ovld __cnfn convert_int8_rte(uchar8);
-int8 __ovld __cnfn convert_int8_sat_rte(uchar8);
-int8 __ovld __cnfn convert_int8_rtz(uchar8);
-int8 __ovld __cnfn convert_int8_sat_rtz(uchar8);
-int8 __ovld __cnfn convert_int8_rtp(uchar8);
-int8 __ovld __cnfn convert_int8_sat_rtp(uchar8);
-int8 __ovld __cnfn convert_int8_rtn(uchar8);
-int8 __ovld __cnfn convert_int8_sat_rtn(uchar8);
-int8 __ovld __cnfn convert_int8(uchar8);
-int8 __ovld __cnfn convert_int8_sat(uchar8);
-int8 __ovld __cnfn convert_int8_rte(short8);
-int8 __ovld __cnfn convert_int8_sat_rte(short8);
-int8 __ovld __cnfn convert_int8_rtz(short8);
-int8 __ovld __cnfn convert_int8_sat_rtz(short8);
-int8 __ovld __cnfn convert_int8_rtp(short8);
-int8 __ovld __cnfn convert_int8_sat_rtp(short8);
-int8 __ovld __cnfn convert_int8_rtn(short8);
-int8 __ovld __cnfn convert_int8_sat_rtn(short8);
-int8 __ovld __cnfn convert_int8(short8);
-int8 __ovld __cnfn convert_int8_sat(short8);
-int8 __ovld __cnfn convert_int8_rte(ushort8);
-int8 __ovld __cnfn convert_int8_sat_rte(ushort8);
-int8 __ovld __cnfn convert_int8_rtz(ushort8);
-int8 __ovld __cnfn convert_int8_sat_rtz(ushort8);
-int8 __ovld __cnfn convert_int8_rtp(ushort8);
-int8 __ovld __cnfn convert_int8_sat_rtp(ushort8);
-int8 __ovld __cnfn convert_int8_rtn(ushort8);
-int8 __ovld __cnfn convert_int8_sat_rtn(ushort8);
-int8 __ovld __cnfn convert_int8(ushort8);
-int8 __ovld __cnfn convert_int8_sat(ushort8);
-int8 __ovld __cnfn convert_int8_rte(int8);
-int8 __ovld __cnfn convert_int8_sat_rte(int8);
-int8 __ovld __cnfn convert_int8_rtz(int8);
-int8 __ovld __cnfn convert_int8_sat_rtz(int8);
-int8 __ovld __cnfn convert_int8_rtp(int8);
-int8 __ovld __cnfn convert_int8_sat_rtp(int8);
-int8 __ovld __cnfn convert_int8_rtn(int8);
-int8 __ovld __cnfn convert_int8_sat_rtn(int8);
-int8 __ovld __cnfn convert_int8(int8);
-int8 __ovld __cnfn convert_int8_sat(int8);
-int8 __ovld __cnfn convert_int8_rte(uint8);
-int8 __ovld __cnfn convert_int8_sat_rte(uint8);
-int8 __ovld __cnfn convert_int8_rtz(uint8);
-int8 __ovld __cnfn convert_int8_sat_rtz(uint8);
-int8 __ovld __cnfn convert_int8_rtp(uint8);
-int8 __ovld __cnfn convert_int8_sat_rtp(uint8);
-int8 __ovld __cnfn convert_int8_rtn(uint8);
-int8 __ovld __cnfn convert_int8_sat_rtn(uint8);
-int8 __ovld __cnfn convert_int8(uint8);
-int8 __ovld __cnfn convert_int8_sat(uint8);
-int8 __ovld __cnfn convert_int8_rte(long8);
-int8 __ovld __cnfn convert_int8_sat_rte(long8);
-int8 __ovld __cnfn convert_int8_rtz(long8);
-int8 __ovld __cnfn convert_int8_sat_rtz(long8);
-int8 __ovld __cnfn convert_int8_rtp(long8);
-int8 __ovld __cnfn convert_int8_sat_rtp(long8);
-int8 __ovld __cnfn convert_int8_rtn(long8);
-int8 __ovld __cnfn convert_int8_sat_rtn(long8);
-int8 __ovld __cnfn convert_int8(long8);
-int8 __ovld __cnfn convert_int8_sat(long8);
-int8 __ovld __cnfn convert_int8_rte(ulong8);
-int8 __ovld __cnfn convert_int8_sat_rte(ulong8);
-int8 __ovld __cnfn convert_int8_rtz(ulong8);
-int8 __ovld __cnfn convert_int8_sat_rtz(ulong8);
-int8 __ovld __cnfn convert_int8_rtp(ulong8);
-int8 __ovld __cnfn convert_int8_sat_rtp(ulong8);
-int8 __ovld __cnfn convert_int8_rtn(ulong8);
-int8 __ovld __cnfn convert_int8_sat_rtn(ulong8);
-int8 __ovld __cnfn convert_int8(ulong8);
-int8 __ovld __cnfn convert_int8_sat(ulong8);
-int8 __ovld __cnfn convert_int8_rte(float8);
-int8 __ovld __cnfn convert_int8_sat_rte(float8);
-int8 __ovld __cnfn convert_int8_rtz(float8);
-int8 __ovld __cnfn convert_int8_sat_rtz(float8);
-int8 __ovld __cnfn convert_int8_rtp(float8);
-int8 __ovld __cnfn convert_int8_sat_rtp(float8);
-int8 __ovld __cnfn convert_int8_rtn(float8);
-int8 __ovld __cnfn convert_int8_sat_rtn(float8);
-int8 __ovld __cnfn convert_int8(float8);
-int8 __ovld __cnfn convert_int8_sat(float8);
-uint8 __ovld __cnfn convert_uint8_rte(char8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(char8);
-uint8 __ovld __cnfn convert_uint8_rtz(char8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(char8);
-uint8 __ovld __cnfn convert_uint8_rtp(char8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(char8);
-uint8 __ovld __cnfn convert_uint8_rtn(char8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(char8);
-uint8 __ovld __cnfn convert_uint8(char8);
-uint8 __ovld __cnfn convert_uint8_sat(char8);
-uint8 __ovld __cnfn convert_uint8_rte(uchar8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(uchar8);
-uint8 __ovld __cnfn convert_uint8_rtz(uchar8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(uchar8);
-uint8 __ovld __cnfn convert_uint8_rtp(uchar8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(uchar8);
-uint8 __ovld __cnfn convert_uint8_rtn(uchar8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(uchar8);
-uint8 __ovld __cnfn convert_uint8(uchar8);
-uint8 __ovld __cnfn convert_uint8_sat(uchar8);
-uint8 __ovld __cnfn convert_uint8_rte(short8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(short8);
-uint8 __ovld __cnfn convert_uint8_rtz(short8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(short8);
-uint8 __ovld __cnfn convert_uint8_rtp(short8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(short8);
-uint8 __ovld __cnfn convert_uint8_rtn(short8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(short8);
-uint8 __ovld __cnfn convert_uint8(short8);
-uint8 __ovld __cnfn convert_uint8_sat(short8);
-uint8 __ovld __cnfn convert_uint8_rte(ushort8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(ushort8);
-uint8 __ovld __cnfn convert_uint8_rtz(ushort8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(ushort8);
-uint8 __ovld __cnfn convert_uint8_rtp(ushort8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(ushort8);
-uint8 __ovld __cnfn convert_uint8_rtn(ushort8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(ushort8);
-uint8 __ovld __cnfn convert_uint8(ushort8);
-uint8 __ovld __cnfn convert_uint8_sat(ushort8);
-uint8 __ovld __cnfn convert_uint8_rte(int8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(int8);
-uint8 __ovld __cnfn convert_uint8_rtz(int8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(int8);
-uint8 __ovld __cnfn convert_uint8_rtp(int8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(int8);
-uint8 __ovld __cnfn convert_uint8_rtn(int8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(int8);
-uint8 __ovld __cnfn convert_uint8(int8);
-uint8 __ovld __cnfn convert_uint8_sat(int8);
-uint8 __ovld __cnfn convert_uint8_rte(uint8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(uint8);
-uint8 __ovld __cnfn convert_uint8_rtz(uint8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(uint8);
-uint8 __ovld __cnfn convert_uint8_rtp(uint8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(uint8);
-uint8 __ovld __cnfn convert_uint8_rtn(uint8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(uint8);
-uint8 __ovld __cnfn convert_uint8(uint8);
-uint8 __ovld __cnfn convert_uint8_sat(uint8);
-uint8 __ovld __cnfn convert_uint8_rte(long8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(long8);
-uint8 __ovld __cnfn convert_uint8_rtz(long8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(long8);
-uint8 __ovld __cnfn convert_uint8_rtp(long8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(long8);
-uint8 __ovld __cnfn convert_uint8_rtn(long8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(long8);
-uint8 __ovld __cnfn convert_uint8(long8);
-uint8 __ovld __cnfn convert_uint8_sat(long8);
-uint8 __ovld __cnfn convert_uint8_rte(ulong8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(ulong8);
-uint8 __ovld __cnfn convert_uint8_rtz(ulong8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(ulong8);
-uint8 __ovld __cnfn convert_uint8_rtp(ulong8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(ulong8);
-uint8 __ovld __cnfn convert_uint8_rtn(ulong8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(ulong8);
-uint8 __ovld __cnfn convert_uint8(ulong8);
-uint8 __ovld __cnfn convert_uint8_sat(ulong8);
-uint8 __ovld __cnfn convert_uint8_rte(float8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(float8);
-uint8 __ovld __cnfn convert_uint8_rtz(float8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(float8);
-uint8 __ovld __cnfn convert_uint8_rtp(float8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(float8);
-uint8 __ovld __cnfn convert_uint8_rtn(float8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(float8);
-uint8 __ovld __cnfn convert_uint8(float8);
-uint8 __ovld __cnfn convert_uint8_sat(float8);
-long8 __ovld __cnfn convert_long8_rte(char8);
-long8 __ovld __cnfn convert_long8_sat_rte(char8);
-long8 __ovld __cnfn convert_long8_rtz(char8);
-long8 __ovld __cnfn convert_long8_sat_rtz(char8);
-long8 __ovld __cnfn convert_long8_rtp(char8);
-long8 __ovld __cnfn convert_long8_sat_rtp(char8);
-long8 __ovld __cnfn convert_long8_rtn(char8);
-long8 __ovld __cnfn convert_long8_sat_rtn(char8);
-long8 __ovld __cnfn convert_long8(char8);
-long8 __ovld __cnfn convert_long8_sat(char8);
-long8 __ovld __cnfn convert_long8_rte(uchar8);
-long8 __ovld __cnfn convert_long8_sat_rte(uchar8);
-long8 __ovld __cnfn convert_long8_rtz(uchar8);
-long8 __ovld __cnfn convert_long8_sat_rtz(uchar8);
-long8 __ovld __cnfn convert_long8_rtp(uchar8);
-long8 __ovld __cnfn convert_long8_sat_rtp(uchar8);
-long8 __ovld __cnfn convert_long8_rtn(uchar8);
-long8 __ovld __cnfn convert_long8_sat_rtn(uchar8);
-long8 __ovld __cnfn convert_long8(uchar8);
-long8 __ovld __cnfn convert_long8_sat(uchar8);
-long8 __ovld __cnfn convert_long8_rte(short8);
-long8 __ovld __cnfn convert_long8_sat_rte(short8);
-long8 __ovld __cnfn convert_long8_rtz(short8);
-long8 __ovld __cnfn convert_long8_sat_rtz(short8);
-long8 __ovld __cnfn convert_long8_rtp(short8);
-long8 __ovld __cnfn convert_long8_sat_rtp(short8);
-long8 __ovld __cnfn convert_long8_rtn(short8);
-long8 __ovld __cnfn convert_long8_sat_rtn(short8);
-long8 __ovld __cnfn convert_long8(short8);
-long8 __ovld __cnfn convert_long8_sat(short8);
-long8 __ovld __cnfn convert_long8_rte(ushort8);
-long8 __ovld __cnfn convert_long8_sat_rte(ushort8);
-long8 __ovld __cnfn convert_long8_rtz(ushort8);
-long8 __ovld __cnfn convert_long8_sat_rtz(ushort8);
-long8 __ovld __cnfn convert_long8_rtp(ushort8);
-long8 __ovld __cnfn convert_long8_sat_rtp(ushort8);
-long8 __ovld __cnfn convert_long8_rtn(ushort8);
-long8 __ovld __cnfn convert_long8_sat_rtn(ushort8);
-long8 __ovld __cnfn convert_long8(ushort8);
-long8 __ovld __cnfn convert_long8_sat(ushort8);
-long8 __ovld __cnfn convert_long8_rte(int8);
-long8 __ovld __cnfn convert_long8_sat_rte(int8);
-long8 __ovld __cnfn convert_long8_rtz(int8);
-long8 __ovld __cnfn convert_long8_sat_rtz(int8);
-long8 __ovld __cnfn convert_long8_rtp(int8);
-long8 __ovld __cnfn convert_long8_sat_rtp(int8);
-long8 __ovld __cnfn convert_long8_rtn(int8);
-long8 __ovld __cnfn convert_long8_sat_rtn(int8);
-long8 __ovld __cnfn convert_long8(int8);
-long8 __ovld __cnfn convert_long8_sat(int8);
-long8 __ovld __cnfn convert_long8_rte(uint8);
-long8 __ovld __cnfn convert_long8_sat_rte(uint8);
-long8 __ovld __cnfn convert_long8_rtz(uint8);
-long8 __ovld __cnfn convert_long8_sat_rtz(uint8);
-long8 __ovld __cnfn convert_long8_rtp(uint8);
-long8 __ovld __cnfn convert_long8_sat_rtp(uint8);
-long8 __ovld __cnfn convert_long8_rtn(uint8);
-long8 __ovld __cnfn convert_long8_sat_rtn(uint8);
-long8 __ovld __cnfn convert_long8(uint8);
-long8 __ovld __cnfn convert_long8_sat(uint8);
-long8 __ovld __cnfn convert_long8_rte(long8);
-long8 __ovld __cnfn convert_long8_sat_rte(long8);
-long8 __ovld __cnfn convert_long8_rtz(long8);
-long8 __ovld __cnfn convert_long8_sat_rtz(long8);
-long8 __ovld __cnfn convert_long8_rtp(long8);
-long8 __ovld __cnfn convert_long8_sat_rtp(long8);
-long8 __ovld __cnfn convert_long8_rtn(long8);
-long8 __ovld __cnfn convert_long8_sat_rtn(long8);
-long8 __ovld __cnfn convert_long8(long8);
-long8 __ovld __cnfn convert_long8_sat(long8);
-long8 __ovld __cnfn convert_long8_rte(ulong8);
-long8 __ovld __cnfn convert_long8_sat_rte(ulong8);
-long8 __ovld __cnfn convert_long8_rtz(ulong8);
-long8 __ovld __cnfn convert_long8_sat_rtz(ulong8);
-long8 __ovld __cnfn convert_long8_rtp(ulong8);
-long8 __ovld __cnfn convert_long8_sat_rtp(ulong8);
-long8 __ovld __cnfn convert_long8_rtn(ulong8);
-long8 __ovld __cnfn convert_long8_sat_rtn(ulong8);
-long8 __ovld __cnfn convert_long8(ulong8);
-long8 __ovld __cnfn convert_long8_sat(ulong8);
-long8 __ovld __cnfn convert_long8_rte(float8);
-long8 __ovld __cnfn convert_long8_sat_rte(float8);
-long8 __ovld __cnfn convert_long8_rtz(float8);
-long8 __ovld __cnfn convert_long8_sat_rtz(float8);
-long8 __ovld __cnfn convert_long8_rtp(float8);
-long8 __ovld __cnfn convert_long8_sat_rtp(float8);
-long8 __ovld __cnfn convert_long8_rtn(float8);
-long8 __ovld __cnfn convert_long8_sat_rtn(float8);
-long8 __ovld __cnfn convert_long8(float8);
-long8 __ovld __cnfn convert_long8_sat(float8);
-ulong8 __ovld __cnfn convert_ulong8_rte(char8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(char8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(char8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(char8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(char8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(char8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(char8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(char8);
-ulong8 __ovld __cnfn convert_ulong8(char8);
-ulong8 __ovld __cnfn convert_ulong8_sat(char8);
-ulong8 __ovld __cnfn convert_ulong8_rte(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(uchar8);
-ulong8 __ovld __cnfn convert_ulong8(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_sat(uchar8);
-ulong8 __ovld __cnfn convert_ulong8_rte(short8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(short8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(short8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(short8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(short8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(short8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(short8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(short8);
-ulong8 __ovld __cnfn convert_ulong8(short8);
-ulong8 __ovld __cnfn convert_ulong8_sat(short8);
-ulong8 __ovld __cnfn convert_ulong8_rte(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(ushort8);
-ulong8 __ovld __cnfn convert_ulong8(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_sat(ushort8);
-ulong8 __ovld __cnfn convert_ulong8_rte(int8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(int8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(int8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(int8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(int8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(int8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(int8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(int8);
-ulong8 __ovld __cnfn convert_ulong8(int8);
-ulong8 __ovld __cnfn convert_ulong8_sat(int8);
-ulong8 __ovld __cnfn convert_ulong8_rte(uint8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(uint8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(uint8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(uint8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(uint8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(uint8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(uint8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(uint8);
-ulong8 __ovld __cnfn convert_ulong8(uint8);
-ulong8 __ovld __cnfn convert_ulong8_sat(uint8);
-ulong8 __ovld __cnfn convert_ulong8_rte(long8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(long8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(long8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(long8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(long8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(long8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(long8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(long8);
-ulong8 __ovld __cnfn convert_ulong8(long8);
-ulong8 __ovld __cnfn convert_ulong8_sat(long8);
-ulong8 __ovld __cnfn convert_ulong8_rte(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(ulong8);
-ulong8 __ovld __cnfn convert_ulong8(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_sat(ulong8);
-ulong8 __ovld __cnfn convert_ulong8_rte(float8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(float8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(float8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(float8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(float8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(float8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(float8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(float8);
-ulong8 __ovld __cnfn convert_ulong8(float8);
-ulong8 __ovld __cnfn convert_ulong8_sat(float8);
-float8 __ovld __cnfn convert_float8_rte(char8);
-float8 __ovld __cnfn convert_float8_rtz(char8);
-float8 __ovld __cnfn convert_float8_rtp(char8);
-float8 __ovld __cnfn convert_float8_rtn(char8);
-float8 __ovld __cnfn convert_float8(char8);
-float8 __ovld __cnfn convert_float8_rte(uchar8);
-float8 __ovld __cnfn convert_float8_rtz(uchar8);
-float8 __ovld __cnfn convert_float8_rtp(uchar8);
-float8 __ovld __cnfn convert_float8_rtn(uchar8);
-float8 __ovld __cnfn convert_float8(uchar8);
-float8 __ovld __cnfn convert_float8_rte(short8);
-float8 __ovld __cnfn convert_float8_rtz(short8);
-float8 __ovld __cnfn convert_float8_rtp(short8);
-float8 __ovld __cnfn convert_float8_rtn(short8);
-float8 __ovld __cnfn convert_float8(short8);
-float8 __ovld __cnfn convert_float8_rte(ushort8);
-float8 __ovld __cnfn convert_float8_rtz(ushort8);
-float8 __ovld __cnfn convert_float8_rtp(ushort8);
-float8 __ovld __cnfn convert_float8_rtn(ushort8);
-float8 __ovld __cnfn convert_float8(ushort8);
-float8 __ovld __cnfn convert_float8_rte(int8);
-float8 __ovld __cnfn convert_float8_rtz(int8);
-float8 __ovld __cnfn convert_float8_rtp(int8);
-float8 __ovld __cnfn convert_float8_rtn(int8);
-float8 __ovld __cnfn convert_float8(int8);
-float8 __ovld __cnfn convert_float8_rte(uint8);
-float8 __ovld __cnfn convert_float8_rtz(uint8);
-float8 __ovld __cnfn convert_float8_rtp(uint8);
-float8 __ovld __cnfn convert_float8_rtn(uint8);
-float8 __ovld __cnfn convert_float8(uint8);
-float8 __ovld __cnfn convert_float8_rte(long8);
-float8 __ovld __cnfn convert_float8_rtz(long8);
-float8 __ovld __cnfn convert_float8_rtp(long8);
-float8 __ovld __cnfn convert_float8_rtn(long8);
-float8 __ovld __cnfn convert_float8(long8);
-float8 __ovld __cnfn convert_float8_rte(ulong8);
-float8 __ovld __cnfn convert_float8_rtz(ulong8);
-float8 __ovld __cnfn convert_float8_rtp(ulong8);
-float8 __ovld __cnfn convert_float8_rtn(ulong8);
-float8 __ovld __cnfn convert_float8(ulong8);
-float8 __ovld __cnfn convert_float8_rte(float8);
-float8 __ovld __cnfn convert_float8_rtz(float8);
-float8 __ovld __cnfn convert_float8_rtp(float8);
-float8 __ovld __cnfn convert_float8_rtn(float8);
-float8 __ovld __cnfn convert_float8(float8);
-char16 __ovld __cnfn convert_char16_rte(char16);
-char16 __ovld __cnfn convert_char16_sat_rte(char16);
-char16 __ovld __cnfn convert_char16_rtz(char16);
-char16 __ovld __cnfn convert_char16_sat_rtz(char16);
-char16 __ovld __cnfn convert_char16_rtp(char16);
-char16 __ovld __cnfn convert_char16_sat_rtp(char16);
-char16 __ovld __cnfn convert_char16_rtn(char16);
-char16 __ovld __cnfn convert_char16_sat_rtn(char16);
-char16 __ovld __cnfn convert_char16(char16);
-char16 __ovld __cnfn convert_char16_sat(char16);
-char16 __ovld __cnfn convert_char16_rte(uchar16);
-char16 __ovld __cnfn convert_char16_sat_rte(uchar16);
-char16 __ovld __cnfn convert_char16_rtz(uchar16);
-char16 __ovld __cnfn convert_char16_sat_rtz(uchar16);
-char16 __ovld __cnfn convert_char16_rtp(uchar16);
-char16 __ovld __cnfn convert_char16_sat_rtp(uchar16);
-char16 __ovld __cnfn convert_char16_rtn(uchar16);
-char16 __ovld __cnfn convert_char16_sat_rtn(uchar16);
-char16 __ovld __cnfn convert_char16(uchar16);
-char16 __ovld __cnfn convert_char16_sat(uchar16);
-char16 __ovld __cnfn convert_char16_rte(short16);
-char16 __ovld __cnfn convert_char16_sat_rte(short16);
-char16 __ovld __cnfn convert_char16_rtz(short16);
-char16 __ovld __cnfn convert_char16_sat_rtz(short16);
-char16 __ovld __cnfn convert_char16_rtp(short16);
-char16 __ovld __cnfn convert_char16_sat_rtp(short16);
-char16 __ovld __cnfn convert_char16_rtn(short16);
-char16 __ovld __cnfn convert_char16_sat_rtn(short16);
-char16 __ovld __cnfn convert_char16(short16);
-char16 __ovld __cnfn convert_char16_sat(short16);
-char16 __ovld __cnfn convert_char16_rte(ushort16);
-char16 __ovld __cnfn convert_char16_sat_rte(ushort16);
-char16 __ovld __cnfn convert_char16_rtz(ushort16);
-char16 __ovld __cnfn convert_char16_sat_rtz(ushort16);
-char16 __ovld __cnfn convert_char16_rtp(ushort16);
-char16 __ovld __cnfn convert_char16_sat_rtp(ushort16);
-char16 __ovld __cnfn convert_char16_rtn(ushort16);
-char16 __ovld __cnfn convert_char16_sat_rtn(ushort16);
-char16 __ovld __cnfn convert_char16(ushort16);
-char16 __ovld __cnfn convert_char16_sat(ushort16);
-char16 __ovld __cnfn convert_char16_rte(int16);
-char16 __ovld __cnfn convert_char16_sat_rte(int16);
-char16 __ovld __cnfn convert_char16_rtz(int16);
-char16 __ovld __cnfn convert_char16_sat_rtz(int16);
-char16 __ovld __cnfn convert_char16_rtp(int16);
-char16 __ovld __cnfn convert_char16_sat_rtp(int16);
-char16 __ovld __cnfn convert_char16_rtn(int16);
-char16 __ovld __cnfn convert_char16_sat_rtn(int16);
-char16 __ovld __cnfn convert_char16(int16);
-char16 __ovld __cnfn convert_char16_sat(int16);
-char16 __ovld __cnfn convert_char16_rte(uint16);
-char16 __ovld __cnfn convert_char16_sat_rte(uint16);
-char16 __ovld __cnfn convert_char16_rtz(uint16);
-char16 __ovld __cnfn convert_char16_sat_rtz(uint16);
-char16 __ovld __cnfn convert_char16_rtp(uint16);
-char16 __ovld __cnfn convert_char16_sat_rtp(uint16);
-char16 __ovld __cnfn convert_char16_rtn(uint16);
-char16 __ovld __cnfn convert_char16_sat_rtn(uint16);
-char16 __ovld __cnfn convert_char16(uint16);
-char16 __ovld __cnfn convert_char16_sat(uint16);
-char16 __ovld __cnfn convert_char16_rte(long16);
-char16 __ovld __cnfn convert_char16_sat_rte(long16);
-char16 __ovld __cnfn convert_char16_rtz(long16);
-char16 __ovld __cnfn convert_char16_sat_rtz(long16);
-char16 __ovld __cnfn convert_char16_rtp(long16);
-char16 __ovld __cnfn convert_char16_sat_rtp(long16);
-char16 __ovld __cnfn convert_char16_rtn(long16);
-char16 __ovld __cnfn convert_char16_sat_rtn(long16);
-char16 __ovld __cnfn convert_char16(long16);
-char16 __ovld __cnfn convert_char16_sat(long16);
-char16 __ovld __cnfn convert_char16_rte(ulong16);
-char16 __ovld __cnfn convert_char16_sat_rte(ulong16);
-char16 __ovld __cnfn convert_char16_rtz(ulong16);
-char16 __ovld __cnfn convert_char16_sat_rtz(ulong16);
-char16 __ovld __cnfn convert_char16_rtp(ulong16);
-char16 __ovld __cnfn convert_char16_sat_rtp(ulong16);
-char16 __ovld __cnfn convert_char16_rtn(ulong16);
-char16 __ovld __cnfn convert_char16_sat_rtn(ulong16);
-char16 __ovld __cnfn convert_char16(ulong16);
-char16 __ovld __cnfn convert_char16_sat(ulong16);
-char16 __ovld __cnfn convert_char16_rte(float16);
-char16 __ovld __cnfn convert_char16_sat_rte(float16);
-char16 __ovld __cnfn convert_char16_rtz(float16);
-char16 __ovld __cnfn convert_char16_sat_rtz(float16);
-char16 __ovld __cnfn convert_char16_rtp(float16);
-char16 __ovld __cnfn convert_char16_sat_rtp(float16);
-char16 __ovld __cnfn convert_char16_rtn(float16);
-char16 __ovld __cnfn convert_char16_sat_rtn(float16);
-char16 __ovld __cnfn convert_char16(float16);
-char16 __ovld __cnfn convert_char16_sat(float16);
-uchar16 __ovld __cnfn convert_uchar16_rte(char16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(char16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(char16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(char16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(char16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(char16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(char16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(char16);
-uchar16 __ovld __cnfn convert_uchar16(char16);
-uchar16 __ovld __cnfn convert_uchar16_sat(char16);
-uchar16 __ovld __cnfn convert_uchar16_rte(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(uchar16);
-uchar16 __ovld __cnfn convert_uchar16(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_sat(uchar16);
-uchar16 __ovld __cnfn convert_uchar16_rte(short16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(short16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(short16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(short16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(short16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(short16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(short16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(short16);
-uchar16 __ovld __cnfn convert_uchar16(short16);
-uchar16 __ovld __cnfn convert_uchar16_sat(short16);
-uchar16 __ovld __cnfn convert_uchar16_rte(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(ushort16);
-uchar16 __ovld __cnfn convert_uchar16(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_sat(ushort16);
-uchar16 __ovld __cnfn convert_uchar16_rte(int16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(int16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(int16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(int16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(int16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(int16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(int16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(int16);
-uchar16 __ovld __cnfn convert_uchar16(int16);
-uchar16 __ovld __cnfn convert_uchar16_sat(int16);
-uchar16 __ovld __cnfn convert_uchar16_rte(uint16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(uint16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(uint16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(uint16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(uint16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(uint16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(uint16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(uint16);
-uchar16 __ovld __cnfn convert_uchar16(uint16);
-uchar16 __ovld __cnfn convert_uchar16_sat(uint16);
-uchar16 __ovld __cnfn convert_uchar16_rte(long16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(long16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(long16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(long16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(long16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(long16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(long16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(long16);
-uchar16 __ovld __cnfn convert_uchar16(long16);
-uchar16 __ovld __cnfn convert_uchar16_sat(long16);
-uchar16 __ovld __cnfn convert_uchar16_rte(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(ulong16);
-uchar16 __ovld __cnfn convert_uchar16(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_sat(ulong16);
-uchar16 __ovld __cnfn convert_uchar16_rte(float16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(float16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(float16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(float16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(float16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(float16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(float16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(float16);
-uchar16 __ovld __cnfn convert_uchar16(float16);
-uchar16 __ovld __cnfn convert_uchar16_sat(float16);
-short16 __ovld __cnfn convert_short16_rte(char16);
-short16 __ovld __cnfn convert_short16_sat_rte(char16);
-short16 __ovld __cnfn convert_short16_rtz(char16);
-short16 __ovld __cnfn convert_short16_sat_rtz(char16);
-short16 __ovld __cnfn convert_short16_rtp(char16);
-short16 __ovld __cnfn convert_short16_sat_rtp(char16);
-short16 __ovld __cnfn convert_short16_rtn(char16);
-short16 __ovld __cnfn convert_short16_sat_rtn(char16);
-short16 __ovld __cnfn convert_short16(char16);
-short16 __ovld __cnfn convert_short16_sat(char16);
-short16 __ovld __cnfn convert_short16_rte(uchar16);
-short16 __ovld __cnfn convert_short16_sat_rte(uchar16);
-short16 __ovld __cnfn convert_short16_rtz(uchar16);
-short16 __ovld __cnfn convert_short16_sat_rtz(uchar16);
-short16 __ovld __cnfn convert_short16_rtp(uchar16);
-short16 __ovld __cnfn convert_short16_sat_rtp(uchar16);
-short16 __ovld __cnfn convert_short16_rtn(uchar16);
-short16 __ovld __cnfn convert_short16_sat_rtn(uchar16);
-short16 __ovld __cnfn convert_short16(uchar16);
-short16 __ovld __cnfn convert_short16_sat(uchar16);
-short16 __ovld __cnfn convert_short16_rte(short16);
-short16 __ovld __cnfn convert_short16_sat_rte(short16);
-short16 __ovld __cnfn convert_short16_rtz(short16);
-short16 __ovld __cnfn convert_short16_sat_rtz(short16);
-short16 __ovld __cnfn convert_short16_rtp(short16);
-short16 __ovld __cnfn convert_short16_sat_rtp(short16);
-short16 __ovld __cnfn convert_short16_rtn(short16);
-short16 __ovld __cnfn convert_short16_sat_rtn(short16);
-short16 __ovld __cnfn convert_short16(short16);
-short16 __ovld __cnfn convert_short16_sat(short16);
-short16 __ovld __cnfn convert_short16_rte(ushort16);
-short16 __ovld __cnfn convert_short16_sat_rte(ushort16);
-short16 __ovld __cnfn convert_short16_rtz(ushort16);
-short16 __ovld __cnfn convert_short16_sat_rtz(ushort16);
-short16 __ovld __cnfn convert_short16_rtp(ushort16);
-short16 __ovld __cnfn convert_short16_sat_rtp(ushort16);
-short16 __ovld __cnfn convert_short16_rtn(ushort16);
-short16 __ovld __cnfn convert_short16_sat_rtn(ushort16);
-short16 __ovld __cnfn convert_short16(ushort16);
-short16 __ovld __cnfn convert_short16_sat(ushort16);
-short16 __ovld __cnfn convert_short16_rte(int16);
-short16 __ovld __cnfn convert_short16_sat_rte(int16);
-short16 __ovld __cnfn convert_short16_rtz(int16);
-short16 __ovld __cnfn convert_short16_sat_rtz(int16);
-short16 __ovld __cnfn convert_short16_rtp(int16);
-short16 __ovld __cnfn convert_short16_sat_rtp(int16);
-short16 __ovld __cnfn convert_short16_rtn(int16);
-short16 __ovld __cnfn convert_short16_sat_rtn(int16);
-short16 __ovld __cnfn convert_short16(int16);
-short16 __ovld __cnfn convert_short16_sat(int16);
-short16 __ovld __cnfn convert_short16_rte(uint16);
-short16 __ovld __cnfn convert_short16_sat_rte(uint16);
-short16 __ovld __cnfn convert_short16_rtz(uint16);
-short16 __ovld __cnfn convert_short16_sat_rtz(uint16);
-short16 __ovld __cnfn convert_short16_rtp(uint16);
-short16 __ovld __cnfn convert_short16_sat_rtp(uint16);
-short16 __ovld __cnfn convert_short16_rtn(uint16);
-short16 __ovld __cnfn convert_short16_sat_rtn(uint16);
-short16 __ovld __cnfn convert_short16(uint16);
-short16 __ovld __cnfn convert_short16_sat(uint16);
-short16 __ovld __cnfn convert_short16_rte(long16);
-short16 __ovld __cnfn convert_short16_sat_rte(long16);
-short16 __ovld __cnfn convert_short16_rtz(long16);
-short16 __ovld __cnfn convert_short16_sat_rtz(long16);
-short16 __ovld __cnfn convert_short16_rtp(long16);
-short16 __ovld __cnfn convert_short16_sat_rtp(long16);
-short16 __ovld __cnfn convert_short16_rtn(long16);
-short16 __ovld __cnfn convert_short16_sat_rtn(long16);
-short16 __ovld __cnfn convert_short16(long16);
-short16 __ovld __cnfn convert_short16_sat(long16);
-short16 __ovld __cnfn convert_short16_rte(ulong16);
-short16 __ovld __cnfn convert_short16_sat_rte(ulong16);
-short16 __ovld __cnfn convert_short16_rtz(ulong16);
-short16 __ovld __cnfn convert_short16_sat_rtz(ulong16);
-short16 __ovld __cnfn convert_short16_rtp(ulong16);
-short16 __ovld __cnfn convert_short16_sat_rtp(ulong16);
-short16 __ovld __cnfn convert_short16_rtn(ulong16);
-short16 __ovld __cnfn convert_short16_sat_rtn(ulong16);
-short16 __ovld __cnfn convert_short16(ulong16);
-short16 __ovld __cnfn convert_short16_sat(ulong16);
-short16 __ovld __cnfn convert_short16_rte(float16);
-short16 __ovld __cnfn convert_short16_sat_rte(float16);
-short16 __ovld __cnfn convert_short16_rtz(float16);
-short16 __ovld __cnfn convert_short16_sat_rtz(float16);
-short16 __ovld __cnfn convert_short16_rtp(float16);
-short16 __ovld __cnfn convert_short16_sat_rtp(float16);
-short16 __ovld __cnfn convert_short16_rtn(float16);
-short16 __ovld __cnfn convert_short16_sat_rtn(float16);
-short16 __ovld __cnfn convert_short16(float16);
-short16 __ovld __cnfn convert_short16_sat(float16);
-ushort16 __ovld __cnfn convert_ushort16_rte(char16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(char16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(char16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(char16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(char16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(char16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(char16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(char16);
-ushort16 __ovld __cnfn convert_ushort16(char16);
-ushort16 __ovld __cnfn convert_ushort16_sat(char16);
-ushort16 __ovld __cnfn convert_ushort16_rte(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(uchar16);
-ushort16 __ovld __cnfn convert_ushort16(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_sat(uchar16);
-ushort16 __ovld __cnfn convert_ushort16_rte(short16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(short16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(short16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(short16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(short16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(short16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(short16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(short16);
-ushort16 __ovld __cnfn convert_ushort16(short16);
-ushort16 __ovld __cnfn convert_ushort16_sat(short16);
-ushort16 __ovld __cnfn convert_ushort16_rte(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(ushort16);
-ushort16 __ovld __cnfn convert_ushort16(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_sat(ushort16);
-ushort16 __ovld __cnfn convert_ushort16_rte(int16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(int16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(int16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(int16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(int16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(int16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(int16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(int16);
-ushort16 __ovld __cnfn convert_ushort16(int16);
-ushort16 __ovld __cnfn convert_ushort16_sat(int16);
-ushort16 __ovld __cnfn convert_ushort16_rte(uint16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(uint16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(uint16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(uint16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(uint16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(uint16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(uint16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(uint16);
-ushort16 __ovld __cnfn convert_ushort16(uint16);
-ushort16 __ovld __cnfn convert_ushort16_sat(uint16);
-ushort16 __ovld __cnfn convert_ushort16_rte(long16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(long16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(long16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(long16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(long16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(long16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(long16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(long16);
-ushort16 __ovld __cnfn convert_ushort16(long16);
-ushort16 __ovld __cnfn convert_ushort16_sat(long16);
-ushort16 __ovld __cnfn convert_ushort16_rte(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(ulong16);
-ushort16 __ovld __cnfn convert_ushort16(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_sat(ulong16);
-ushort16 __ovld __cnfn convert_ushort16_rte(float16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(float16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(float16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(float16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(float16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(float16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(float16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(float16);
-ushort16 __ovld __cnfn convert_ushort16(float16);
-ushort16 __ovld __cnfn convert_ushort16_sat(float16);
-int16 __ovld __cnfn convert_int16_rte(char16);
-int16 __ovld __cnfn convert_int16_sat_rte(char16);
-int16 __ovld __cnfn convert_int16_rtz(char16);
-int16 __ovld __cnfn convert_int16_sat_rtz(char16);
-int16 __ovld __cnfn convert_int16_rtp(char16);
-int16 __ovld __cnfn convert_int16_sat_rtp(char16);
-int16 __ovld __cnfn convert_int16_rtn(char16);
-int16 __ovld __cnfn convert_int16_sat_rtn(char16);
-int16 __ovld __cnfn convert_int16(char16);
-int16 __ovld __cnfn convert_int16_sat(char16);
-int16 __ovld __cnfn convert_int16_rte(uchar16);
-int16 __ovld __cnfn convert_int16_sat_rte(uchar16);
-int16 __ovld __cnfn convert_int16_rtz(uchar16);
-int16 __ovld __cnfn convert_int16_sat_rtz(uchar16);
-int16 __ovld __cnfn convert_int16_rtp(uchar16);
-int16 __ovld __cnfn convert_int16_sat_rtp(uchar16);
-int16 __ovld __cnfn convert_int16_rtn(uchar16);
-int16 __ovld __cnfn convert_int16_sat_rtn(uchar16);
-int16 __ovld __cnfn convert_int16(uchar16);
-int16 __ovld __cnfn convert_int16_sat(uchar16);
-int16 __ovld __cnfn convert_int16_rte(short16);
-int16 __ovld __cnfn convert_int16_sat_rte(short16);
-int16 __ovld __cnfn convert_int16_rtz(short16);
-int16 __ovld __cnfn convert_int16_sat_rtz(short16);
-int16 __ovld __cnfn convert_int16_rtp(short16);
-int16 __ovld __cnfn convert_int16_sat_rtp(short16);
-int16 __ovld __cnfn convert_int16_rtn(short16);
-int16 __ovld __cnfn convert_int16_sat_rtn(short16);
-int16 __ovld __cnfn convert_int16(short16);
-int16 __ovld __cnfn convert_int16_sat(short16);
-int16 __ovld __cnfn convert_int16_rte(ushort16);
-int16 __ovld __cnfn convert_int16_sat_rte(ushort16);
-int16 __ovld __cnfn convert_int16_rtz(ushort16);
-int16 __ovld __cnfn convert_int16_sat_rtz(ushort16);
-int16 __ovld __cnfn convert_int16_rtp(ushort16);
-int16 __ovld __cnfn convert_int16_sat_rtp(ushort16);
-int16 __ovld __cnfn convert_int16_rtn(ushort16);
-int16 __ovld __cnfn convert_int16_sat_rtn(ushort16);
-int16 __ovld __cnfn convert_int16(ushort16);
-int16 __ovld __cnfn convert_int16_sat(ushort16);
-int16 __ovld __cnfn convert_int16_rte(int16);
-int16 __ovld __cnfn convert_int16_sat_rte(int16);
-int16 __ovld __cnfn convert_int16_rtz(int16);
-int16 __ovld __cnfn convert_int16_sat_rtz(int16);
-int16 __ovld __cnfn convert_int16_rtp(int16);
-int16 __ovld __cnfn convert_int16_sat_rtp(int16);
-int16 __ovld __cnfn convert_int16_rtn(int16);
-int16 __ovld __cnfn convert_int16_sat_rtn(int16);
-int16 __ovld __cnfn convert_int16(int16);
-int16 __ovld __cnfn convert_int16_sat(int16);
-int16 __ovld __cnfn convert_int16_rte(uint16);
-int16 __ovld __cnfn convert_int16_sat_rte(uint16);
-int16 __ovld __cnfn convert_int16_rtz(uint16);
-int16 __ovld __cnfn convert_int16_sat_rtz(uint16);
-int16 __ovld __cnfn convert_int16_rtp(uint16);
-int16 __ovld __cnfn convert_int16_sat_rtp(uint16);
-int16 __ovld __cnfn convert_int16_rtn(uint16);
-int16 __ovld __cnfn convert_int16_sat_rtn(uint16);
-int16 __ovld __cnfn convert_int16(uint16);
-int16 __ovld __cnfn convert_int16_sat(uint16);
-int16 __ovld __cnfn convert_int16_rte(long16);
-int16 __ovld __cnfn convert_int16_sat_rte(long16);
-int16 __ovld __cnfn convert_int16_rtz(long16);
-int16 __ovld __cnfn convert_int16_sat_rtz(long16);
-int16 __ovld __cnfn convert_int16_rtp(long16);
-int16 __ovld __cnfn convert_int16_sat_rtp(long16);
-int16 __ovld __cnfn convert_int16_rtn(long16);
-int16 __ovld __cnfn convert_int16_sat_rtn(long16);
-int16 __ovld __cnfn convert_int16(long16);
-int16 __ovld __cnfn convert_int16_sat(long16);
-int16 __ovld __cnfn convert_int16_rte(ulong16);
-int16 __ovld __cnfn convert_int16_sat_rte(ulong16);
-int16 __ovld __cnfn convert_int16_rtz(ulong16);
-int16 __ovld __cnfn convert_int16_sat_rtz(ulong16);
-int16 __ovld __cnfn convert_int16_rtp(ulong16);
-int16 __ovld __cnfn convert_int16_sat_rtp(ulong16);
-int16 __ovld __cnfn convert_int16_rtn(ulong16);
-int16 __ovld __cnfn convert_int16_sat_rtn(ulong16);
-int16 __ovld __cnfn convert_int16(ulong16);
-int16 __ovld __cnfn convert_int16_sat(ulong16);
-int16 __ovld __cnfn convert_int16_rte(float16);
-int16 __ovld __cnfn convert_int16_sat_rte(float16);
-int16 __ovld __cnfn convert_int16_rtz(float16);
-int16 __ovld __cnfn convert_int16_sat_rtz(float16);
-int16 __ovld __cnfn convert_int16_rtp(float16);
-int16 __ovld __cnfn convert_int16_sat_rtp(float16);
-int16 __ovld __cnfn convert_int16_rtn(float16);
-int16 __ovld __cnfn convert_int16_sat_rtn(float16);
-int16 __ovld __cnfn convert_int16(float16);
-int16 __ovld __cnfn convert_int16_sat(float16);
-uint16 __ovld __cnfn convert_uint16_rte(char16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(char16);
-uint16 __ovld __cnfn convert_uint16_rtz(char16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(char16);
-uint16 __ovld __cnfn convert_uint16_rtp(char16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(char16);
-uint16 __ovld __cnfn convert_uint16_rtn(char16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(char16);
-uint16 __ovld __cnfn convert_uint16(char16);
-uint16 __ovld __cnfn convert_uint16_sat(char16);
-uint16 __ovld __cnfn convert_uint16_rte(uchar16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(uchar16);
-uint16 __ovld __cnfn convert_uint16_rtz(uchar16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(uchar16);
-uint16 __ovld __cnfn convert_uint16_rtp(uchar16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(uchar16);
-uint16 __ovld __cnfn convert_uint16_rtn(uchar16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(uchar16);
-uint16 __ovld __cnfn convert_uint16(uchar16);
-uint16 __ovld __cnfn convert_uint16_sat(uchar16);
-uint16 __ovld __cnfn convert_uint16_rte(short16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(short16);
-uint16 __ovld __cnfn convert_uint16_rtz(short16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(short16);
-uint16 __ovld __cnfn convert_uint16_rtp(short16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(short16);
-uint16 __ovld __cnfn convert_uint16_rtn(short16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(short16);
-uint16 __ovld __cnfn convert_uint16(short16);
-uint16 __ovld __cnfn convert_uint16_sat(short16);
-uint16 __ovld __cnfn convert_uint16_rte(ushort16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(ushort16);
-uint16 __ovld __cnfn convert_uint16_rtz(ushort16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(ushort16);
-uint16 __ovld __cnfn convert_uint16_rtp(ushort16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(ushort16);
-uint16 __ovld __cnfn convert_uint16_rtn(ushort16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(ushort16);
-uint16 __ovld __cnfn convert_uint16(ushort16);
-uint16 __ovld __cnfn convert_uint16_sat(ushort16);
-uint16 __ovld __cnfn convert_uint16_rte(int16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(int16);
-uint16 __ovld __cnfn convert_uint16_rtz(int16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(int16);
-uint16 __ovld __cnfn convert_uint16_rtp(int16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(int16);
-uint16 __ovld __cnfn convert_uint16_rtn(int16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(int16);
-uint16 __ovld __cnfn convert_uint16(int16);
-uint16 __ovld __cnfn convert_uint16_sat(int16);
-uint16 __ovld __cnfn convert_uint16_rte(uint16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(uint16);
-uint16 __ovld __cnfn convert_uint16_rtz(uint16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(uint16);
-uint16 __ovld __cnfn convert_uint16_rtp(uint16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(uint16);
-uint16 __ovld __cnfn convert_uint16_rtn(uint16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(uint16);
-uint16 __ovld __cnfn convert_uint16(uint16);
-uint16 __ovld __cnfn convert_uint16_sat(uint16);
-uint16 __ovld __cnfn convert_uint16_rte(long16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(long16);
-uint16 __ovld __cnfn convert_uint16_rtz(long16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(long16);
-uint16 __ovld __cnfn convert_uint16_rtp(long16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(long16);
-uint16 __ovld __cnfn convert_uint16_rtn(long16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(long16);
-uint16 __ovld __cnfn convert_uint16(long16);
-uint16 __ovld __cnfn convert_uint16_sat(long16);
-uint16 __ovld __cnfn convert_uint16_rte(ulong16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(ulong16);
-uint16 __ovld __cnfn convert_uint16_rtz(ulong16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(ulong16);
-uint16 __ovld __cnfn convert_uint16_rtp(ulong16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(ulong16);
-uint16 __ovld __cnfn convert_uint16_rtn(ulong16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(ulong16);
-uint16 __ovld __cnfn convert_uint16(ulong16);
-uint16 __ovld __cnfn convert_uint16_sat(ulong16);
-uint16 __ovld __cnfn convert_uint16_rte(float16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(float16);
-uint16 __ovld __cnfn convert_uint16_rtz(float16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(float16);
-uint16 __ovld __cnfn convert_uint16_rtp(float16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(float16);
-uint16 __ovld __cnfn convert_uint16_rtn(float16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(float16);
-uint16 __ovld __cnfn convert_uint16(float16);
-uint16 __ovld __cnfn convert_uint16_sat(float16);
-long16 __ovld __cnfn convert_long16_rte(char16);
-long16 __ovld __cnfn convert_long16_sat_rte(char16);
-long16 __ovld __cnfn convert_long16_rtz(char16);
-long16 __ovld __cnfn convert_long16_sat_rtz(char16);
-long16 __ovld __cnfn convert_long16_rtp(char16);
-long16 __ovld __cnfn convert_long16_sat_rtp(char16);
-long16 __ovld __cnfn convert_long16_rtn(char16);
-long16 __ovld __cnfn convert_long16_sat_rtn(char16);
-long16 __ovld __cnfn convert_long16(char16);
-long16 __ovld __cnfn convert_long16_sat(char16);
-long16 __ovld __cnfn convert_long16_rte(uchar16);
-long16 __ovld __cnfn convert_long16_sat_rte(uchar16);
-long16 __ovld __cnfn convert_long16_rtz(uchar16);
-long16 __ovld __cnfn convert_long16_sat_rtz(uchar16);
-long16 __ovld __cnfn convert_long16_rtp(uchar16);
-long16 __ovld __cnfn convert_long16_sat_rtp(uchar16);
-long16 __ovld __cnfn convert_long16_rtn(uchar16);
-long16 __ovld __cnfn convert_long16_sat_rtn(uchar16);
-long16 __ovld __cnfn convert_long16(uchar16);
-long16 __ovld __cnfn convert_long16_sat(uchar16);
-long16 __ovld __cnfn convert_long16_rte(short16);
-long16 __ovld __cnfn convert_long16_sat_rte(short16);
-long16 __ovld __cnfn convert_long16_rtz(short16);
-long16 __ovld __cnfn convert_long16_sat_rtz(short16);
-long16 __ovld __cnfn convert_long16_rtp(short16);
-long16 __ovld __cnfn convert_long16_sat_rtp(short16);
-long16 __ovld __cnfn convert_long16_rtn(short16);
-long16 __ovld __cnfn convert_long16_sat_rtn(short16);
-long16 __ovld __cnfn convert_long16(short16);
-long16 __ovld __cnfn convert_long16_sat(short16);
-long16 __ovld __cnfn convert_long16_rte(ushort16);
-long16 __ovld __cnfn convert_long16_sat_rte(ushort16);
-long16 __ovld __cnfn convert_long16_rtz(ushort16);
-long16 __ovld __cnfn convert_long16_sat_rtz(ushort16);
-long16 __ovld __cnfn convert_long16_rtp(ushort16);
-long16 __ovld __cnfn convert_long16_sat_rtp(ushort16);
-long16 __ovld __cnfn convert_long16_rtn(ushort16);
-long16 __ovld __cnfn convert_long16_sat_rtn(ushort16);
-long16 __ovld __cnfn convert_long16(ushort16);
-long16 __ovld __cnfn convert_long16_sat(ushort16);
-long16 __ovld __cnfn convert_long16_rte(int16);
-long16 __ovld __cnfn convert_long16_sat_rte(int16);
-long16 __ovld __cnfn convert_long16_rtz(int16);
-long16 __ovld __cnfn convert_long16_sat_rtz(int16);
-long16 __ovld __cnfn convert_long16_rtp(int16);
-long16 __ovld __cnfn convert_long16_sat_rtp(int16);
-long16 __ovld __cnfn convert_long16_rtn(int16);
-long16 __ovld __cnfn convert_long16_sat_rtn(int16);
-long16 __ovld __cnfn convert_long16(int16);
-long16 __ovld __cnfn convert_long16_sat(int16);
-long16 __ovld __cnfn convert_long16_rte(uint16);
-long16 __ovld __cnfn convert_long16_sat_rte(uint16);
-long16 __ovld __cnfn convert_long16_rtz(uint16);
-long16 __ovld __cnfn convert_long16_sat_rtz(uint16);
-long16 __ovld __cnfn convert_long16_rtp(uint16);
-long16 __ovld __cnfn convert_long16_sat_rtp(uint16);
-long16 __ovld __cnfn convert_long16_rtn(uint16);
-long16 __ovld __cnfn convert_long16_sat_rtn(uint16);
-long16 __ovld __cnfn convert_long16(uint16);
-long16 __ovld __cnfn convert_long16_sat(uint16);
-long16 __ovld __cnfn convert_long16_rte(long16);
-long16 __ovld __cnfn convert_long16_sat_rte(long16);
-long16 __ovld __cnfn convert_long16_rtz(long16);
-long16 __ovld __cnfn convert_long16_sat_rtz(long16);
-long16 __ovld __cnfn convert_long16_rtp(long16);
-long16 __ovld __cnfn convert_long16_sat_rtp(long16);
-long16 __ovld __cnfn convert_long16_rtn(long16);
-long16 __ovld __cnfn convert_long16_sat_rtn(long16);
-long16 __ovld __cnfn convert_long16(long16);
-long16 __ovld __cnfn convert_long16_sat(long16);
-long16 __ovld __cnfn convert_long16_rte(ulong16);
-long16 __ovld __cnfn convert_long16_sat_rte(ulong16);
-long16 __ovld __cnfn convert_long16_rtz(ulong16);
-long16 __ovld __cnfn convert_long16_sat_rtz(ulong16);
-long16 __ovld __cnfn convert_long16_rtp(ulong16);
-long16 __ovld __cnfn convert_long16_sat_rtp(ulong16);
-long16 __ovld __cnfn convert_long16_rtn(ulong16);
-long16 __ovld __cnfn convert_long16_sat_rtn(ulong16);
-long16 __ovld __cnfn convert_long16(ulong16);
-long16 __ovld __cnfn convert_long16_sat(ulong16);
-long16 __ovld __cnfn convert_long16_rte(float16);
-long16 __ovld __cnfn convert_long16_sat_rte(float16);
-long16 __ovld __cnfn convert_long16_rtz(float16);
-long16 __ovld __cnfn convert_long16_sat_rtz(float16);
-long16 __ovld __cnfn convert_long16_rtp(float16);
-long16 __ovld __cnfn convert_long16_sat_rtp(float16);
-long16 __ovld __cnfn convert_long16_rtn(float16);
-long16 __ovld __cnfn convert_long16_sat_rtn(float16);
-long16 __ovld __cnfn convert_long16(float16);
-long16 __ovld __cnfn convert_long16_sat(float16);
-ulong16 __ovld __cnfn convert_ulong16_rte(char16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(char16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(char16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(char16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(char16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(char16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(char16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(char16);
-ulong16 __ovld __cnfn convert_ulong16(char16);
-ulong16 __ovld __cnfn convert_ulong16_sat(char16);
-ulong16 __ovld __cnfn convert_ulong16_rte(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(uchar16);
-ulong16 __ovld __cnfn convert_ulong16(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_sat(uchar16);
-ulong16 __ovld __cnfn convert_ulong16_rte(short16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(short16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(short16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(short16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(short16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(short16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(short16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(short16);
-ulong16 __ovld __cnfn convert_ulong16(short16);
-ulong16 __ovld __cnfn convert_ulong16_sat(short16);
-ulong16 __ovld __cnfn convert_ulong16_rte(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(ushort16);
-ulong16 __ovld __cnfn convert_ulong16(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_sat(ushort16);
-ulong16 __ovld __cnfn convert_ulong16_rte(int16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(int16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(int16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(int16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(int16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(int16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(int16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(int16);
-ulong16 __ovld __cnfn convert_ulong16(int16);
-ulong16 __ovld __cnfn convert_ulong16_sat(int16);
-ulong16 __ovld __cnfn convert_ulong16_rte(uint16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(uint16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(uint16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(uint16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(uint16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(uint16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(uint16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(uint16);
-ulong16 __ovld __cnfn convert_ulong16(uint16);
-ulong16 __ovld __cnfn convert_ulong16_sat(uint16);
-ulong16 __ovld __cnfn convert_ulong16_rte(long16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(long16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(long16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(long16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(long16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(long16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(long16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(long16);
-ulong16 __ovld __cnfn convert_ulong16(long16);
-ulong16 __ovld __cnfn convert_ulong16_sat(long16);
-ulong16 __ovld __cnfn convert_ulong16_rte(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(ulong16);
-ulong16 __ovld __cnfn convert_ulong16(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_sat(ulong16);
-ulong16 __ovld __cnfn convert_ulong16_rte(float16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(float16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(float16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(float16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(float16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(float16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(float16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(float16);
-ulong16 __ovld __cnfn convert_ulong16(float16);
-ulong16 __ovld __cnfn convert_ulong16_sat(float16);
-float16 __ovld __cnfn convert_float16_rte(char16);
-float16 __ovld __cnfn convert_float16_rtz(char16);
-float16 __ovld __cnfn convert_float16_rtp(char16);
-float16 __ovld __cnfn convert_float16_rtn(char16);
-float16 __ovld __cnfn convert_float16(char16);
-float16 __ovld __cnfn convert_float16_rte(uchar16);
-float16 __ovld __cnfn convert_float16_rtz(uchar16);
-float16 __ovld __cnfn convert_float16_rtp(uchar16);
-float16 __ovld __cnfn convert_float16_rtn(uchar16);
-float16 __ovld __cnfn convert_float16(uchar16);
-float16 __ovld __cnfn convert_float16_rte(short16);
-float16 __ovld __cnfn convert_float16_rtz(short16);
-float16 __ovld __cnfn convert_float16_rtp(short16);
-float16 __ovld __cnfn convert_float16_rtn(short16);
-float16 __ovld __cnfn convert_float16(short16);
-float16 __ovld __cnfn convert_float16_rte(ushort16);
-float16 __ovld __cnfn convert_float16_rtz(ushort16);
-float16 __ovld __cnfn convert_float16_rtp(ushort16);
-float16 __ovld __cnfn convert_float16_rtn(ushort16);
-float16 __ovld __cnfn convert_float16(ushort16);
-float16 __ovld __cnfn convert_float16_rte(int16);
-float16 __ovld __cnfn convert_float16_rtz(int16);
-float16 __ovld __cnfn convert_float16_rtp(int16);
-float16 __ovld __cnfn convert_float16_rtn(int16);
-float16 __ovld __cnfn convert_float16(int16);
-float16 __ovld __cnfn convert_float16_rte(uint16);
-float16 __ovld __cnfn convert_float16_rtz(uint16);
-float16 __ovld __cnfn convert_float16_rtp(uint16);
-float16 __ovld __cnfn convert_float16_rtn(uint16);
-float16 __ovld __cnfn convert_float16(uint16);
-float16 __ovld __cnfn convert_float16_rte(long16);
-float16 __ovld __cnfn convert_float16_rtz(long16);
-float16 __ovld __cnfn convert_float16_rtp(long16);
-float16 __ovld __cnfn convert_float16_rtn(long16);
-float16 __ovld __cnfn convert_float16(long16);
-float16 __ovld __cnfn convert_float16_rte(ulong16);
-float16 __ovld __cnfn convert_float16_rtz(ulong16);
-float16 __ovld __cnfn convert_float16_rtp(ulong16);
-float16 __ovld __cnfn convert_float16_rtn(ulong16);
-float16 __ovld __cnfn convert_float16(ulong16);
-float16 __ovld __cnfn convert_float16_rte(float16);
-float16 __ovld __cnfn convert_float16_rtz(float16);
-float16 __ovld __cnfn convert_float16_rtp(float16);
-float16 __ovld __cnfn convert_float16_rtn(float16);
-float16 __ovld __cnfn convert_float16(float16);
-
-// Conversions with double data type parameters or return value.
-
-#ifdef cl_khr_fp64
-char __ovld __cnfn convert_char(double);
-char __ovld __cnfn convert_char_rte(double);
-char __ovld __cnfn convert_char_rtn(double);
-char __ovld __cnfn convert_char_rtp(double);
-char __ovld __cnfn convert_char_rtz(double);
-char __ovld __cnfn convert_char_sat(double);
-char __ovld __cnfn convert_char_sat_rte(double);
-char __ovld __cnfn convert_char_sat_rtn(double);
-char __ovld __cnfn convert_char_sat_rtp(double);
-char __ovld __cnfn convert_char_sat_rtz(double);
-char2 __ovld __cnfn convert_char2(double2);
-char2 __ovld __cnfn convert_char2_rte(double2);
-char2 __ovld __cnfn convert_char2_rtn(double2);
-char2 __ovld __cnfn convert_char2_rtp(double2);
-char2 __ovld __cnfn convert_char2_rtz(double2);
-char2 __ovld __cnfn convert_char2_sat(double2);
-char2 __ovld __cnfn convert_char2_sat_rte(double2);
-char2 __ovld __cnfn convert_char2_sat_rtn(double2);
-char2 __ovld __cnfn convert_char2_sat_rtp(double2);
-char2 __ovld __cnfn convert_char2_sat_rtz(double2);
-char3 __ovld __cnfn convert_char3(double3);
-char3 __ovld __cnfn convert_char3_rte(double3);
-char3 __ovld __cnfn convert_char3_rtn(double3);
-char3 __ovld __cnfn convert_char3_rtp(double3);
-char3 __ovld __cnfn convert_char3_rtz(double3);
-char3 __ovld __cnfn convert_char3_sat(double3);
-char3 __ovld __cnfn convert_char3_sat_rte(double3);
-char3 __ovld __cnfn convert_char3_sat_rtn(double3);
-char3 __ovld __cnfn convert_char3_sat_rtp(double3);
-char3 __ovld __cnfn convert_char3_sat_rtz(double3);
-char4 __ovld __cnfn convert_char4(double4);
-char4 __ovld __cnfn convert_char4_rte(double4);
-char4 __ovld __cnfn convert_char4_rtn(double4);
-char4 __ovld __cnfn convert_char4_rtp(double4);
-char4 __ovld __cnfn convert_char4_rtz(double4);
-char4 __ovld __cnfn convert_char4_sat(double4);
-char4 __ovld __cnfn convert_char4_sat_rte(double4);
-char4 __ovld __cnfn convert_char4_sat_rtn(double4);
-char4 __ovld __cnfn convert_char4_sat_rtp(double4);
-char4 __ovld __cnfn convert_char4_sat_rtz(double4);
-char8 __ovld __cnfn convert_char8(double8);
-char8 __ovld __cnfn convert_char8_rte(double8);
-char8 __ovld __cnfn convert_char8_rtn(double8);
-char8 __ovld __cnfn convert_char8_rtp(double8);
-char8 __ovld __cnfn convert_char8_rtz(double8);
-char8 __ovld __cnfn convert_char8_sat(double8);
-char8 __ovld __cnfn convert_char8_sat_rte(double8);
-char8 __ovld __cnfn convert_char8_sat_rtn(double8);
-char8 __ovld __cnfn convert_char8_sat_rtp(double8);
-char8 __ovld __cnfn convert_char8_sat_rtz(double8);
-char16 __ovld __cnfn convert_char16(double16);
-char16 __ovld __cnfn convert_char16_rte(double16);
-char16 __ovld __cnfn convert_char16_rtn(double16);
-char16 __ovld __cnfn convert_char16_rtp(double16);
-char16 __ovld __cnfn convert_char16_rtz(double16);
-char16 __ovld __cnfn convert_char16_sat(double16);
-char16 __ovld __cnfn convert_char16_sat_rte(double16);
-char16 __ovld __cnfn convert_char16_sat_rtn(double16);
-char16 __ovld __cnfn convert_char16_sat_rtp(double16);
-char16 __ovld __cnfn convert_char16_sat_rtz(double16);
-
-uchar __ovld __cnfn convert_uchar(double);
-uchar __ovld __cnfn convert_uchar_rte(double);
-uchar __ovld __cnfn convert_uchar_rtn(double);
-uchar __ovld __cnfn convert_uchar_rtp(double);
-uchar __ovld __cnfn convert_uchar_rtz(double);
-uchar __ovld __cnfn convert_uchar_sat(double);
-uchar __ovld __cnfn convert_uchar_sat_rte(double);
-uchar __ovld __cnfn convert_uchar_sat_rtn(double);
-uchar __ovld __cnfn convert_uchar_sat_rtp(double);
-uchar __ovld __cnfn convert_uchar_sat_rtz(double);
-uchar2 __ovld __cnfn convert_uchar2(double2);
-uchar2 __ovld __cnfn convert_uchar2_rte(double2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(double2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(double2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(double2);
-uchar2 __ovld __cnfn convert_uchar2_sat(double2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(double2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(double2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(double2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(double2);
-uchar3 __ovld __cnfn convert_uchar3(double3);
-uchar3 __ovld __cnfn convert_uchar3_rte(double3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(double3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(double3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(double3);
-uchar3 __ovld __cnfn convert_uchar3_sat(double3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(double3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(double3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(double3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(double3);
-uchar4 __ovld __cnfn convert_uchar4(double4);
-uchar4 __ovld __cnfn convert_uchar4_rte(double4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(double4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(double4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(double4);
-uchar4 __ovld __cnfn convert_uchar4_sat(double4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(double4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(double4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(double4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(double4);
-uchar8 __ovld __cnfn convert_uchar8(double8);
-uchar8 __ovld __cnfn convert_uchar8_rte(double8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(double8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(double8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(double8);
-uchar8 __ovld __cnfn convert_uchar8_sat(double8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(double8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(double8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(double8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(double8);
-uchar16 __ovld __cnfn convert_uchar16(double16);
-uchar16 __ovld __cnfn convert_uchar16_rte(double16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(double16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(double16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(double16);
-uchar16 __ovld __cnfn convert_uchar16_sat(double16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(double16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(double16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(double16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(double16);
-
-short __ovld __cnfn convert_short(double);
-short __ovld __cnfn convert_short_rte(double);
-short __ovld __cnfn convert_short_rtn(double);
-short __ovld __cnfn convert_short_rtp(double);
-short __ovld __cnfn convert_short_rtz(double);
-short __ovld __cnfn convert_short_sat(double);
-short __ovld __cnfn convert_short_sat_rte(double);
-short __ovld __cnfn convert_short_sat_rtn(double);
-short __ovld __cnfn convert_short_sat_rtp(double);
-short __ovld __cnfn convert_short_sat_rtz(double);
-short2 __ovld __cnfn convert_short2(double2);
-short2 __ovld __cnfn convert_short2_rte(double2);
-short2 __ovld __cnfn convert_short2_rtn(double2);
-short2 __ovld __cnfn convert_short2_rtp(double2);
-short2 __ovld __cnfn convert_short2_rtz(double2);
-short2 __ovld __cnfn convert_short2_sat(double2);
-short2 __ovld __cnfn convert_short2_sat_rte(double2);
-short2 __ovld __cnfn convert_short2_sat_rtn(double2);
-short2 __ovld __cnfn convert_short2_sat_rtp(double2);
-short2 __ovld __cnfn convert_short2_sat_rtz(double2);
-short3 __ovld __cnfn convert_short3(double3);
-short3 __ovld __cnfn convert_short3_rte(double3);
-short3 __ovld __cnfn convert_short3_rtn(double3);
-short3 __ovld __cnfn convert_short3_rtp(double3);
-short3 __ovld __cnfn convert_short3_rtz(double3);
-short3 __ovld __cnfn convert_short3_sat(double3);
-short3 __ovld __cnfn convert_short3_sat_rte(double3);
-short3 __ovld __cnfn convert_short3_sat_rtn(double3);
-short3 __ovld __cnfn convert_short3_sat_rtp(double3);
-short3 __ovld __cnfn convert_short3_sat_rtz(double3);
-short4 __ovld __cnfn convert_short4(double4);
-short4 __ovld __cnfn convert_short4_rte(double4);
-short4 __ovld __cnfn convert_short4_rtn(double4);
-short4 __ovld __cnfn convert_short4_rtp(double4);
-short4 __ovld __cnfn convert_short4_rtz(double4);
-short4 __ovld __cnfn convert_short4_sat(double4);
-short4 __ovld __cnfn convert_short4_sat_rte(double4);
-short4 __ovld __cnfn convert_short4_sat_rtn(double4);
-short4 __ovld __cnfn convert_short4_sat_rtp(double4);
-short4 __ovld __cnfn convert_short4_sat_rtz(double4);
-short8 __ovld __cnfn convert_short8(double8);
-short8 __ovld __cnfn convert_short8_rte(double8);
-short8 __ovld __cnfn convert_short8_rtn(double8);
-short8 __ovld __cnfn convert_short8_rtp(double8);
-short8 __ovld __cnfn convert_short8_rtz(double8);
-short8 __ovld __cnfn convert_short8_sat(double8);
-short8 __ovld __cnfn convert_short8_sat_rte(double8);
-short8 __ovld __cnfn convert_short8_sat_rtn(double8);
-short8 __ovld __cnfn convert_short8_sat_rtp(double8);
-short8 __ovld __cnfn convert_short8_sat_rtz(double8);
-short16 __ovld __cnfn convert_short16(double16);
-short16 __ovld __cnfn convert_short16_rte(double16);
-short16 __ovld __cnfn convert_short16_rtn(double16);
-short16 __ovld __cnfn convert_short16_rtp(double16);
-short16 __ovld __cnfn convert_short16_rtz(double16);
-short16 __ovld __cnfn convert_short16_sat(double16);
-short16 __ovld __cnfn convert_short16_sat_rte(double16);
-short16 __ovld __cnfn convert_short16_sat_rtn(double16);
-short16 __ovld __cnfn convert_short16_sat_rtp(double16);
-short16 __ovld __cnfn convert_short16_sat_rtz(double16);
-
-ushort __ovld __cnfn convert_ushort(double);
-ushort __ovld __cnfn convert_ushort_rte(double);
-ushort __ovld __cnfn convert_ushort_rtn(double);
-ushort __ovld __cnfn convert_ushort_rtp(double);
-ushort __ovld __cnfn convert_ushort_rtz(double);
-ushort __ovld __cnfn convert_ushort_sat(double);
-ushort __ovld __cnfn convert_ushort_sat_rte(double);
-ushort __ovld __cnfn convert_ushort_sat_rtn(double);
-ushort __ovld __cnfn convert_ushort_sat_rtp(double);
-ushort __ovld __cnfn convert_ushort_sat_rtz(double);
-ushort2 __ovld __cnfn convert_ushort2(double2);
-ushort2 __ovld __cnfn convert_ushort2_rte(double2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(double2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(double2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(double2);
-ushort2 __ovld __cnfn convert_ushort2_sat(double2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(double2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(double2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(double2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(double2);
-ushort3 __ovld __cnfn convert_ushort3(double3);
-ushort3 __ovld __cnfn convert_ushort3_rte(double3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(double3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(double3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(double3);
-ushort3 __ovld __cnfn convert_ushort3_sat(double3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(double3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(double3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(double3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(double3);
-ushort4 __ovld __cnfn convert_ushort4(double4);
-ushort4 __ovld __cnfn convert_ushort4_rte(double4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(double4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(double4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(double4);
-ushort4 __ovld __cnfn convert_ushort4_sat(double4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(double4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(double4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(double4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(double4);
-ushort8 __ovld __cnfn convert_ushort8(double8);
-ushort8 __ovld __cnfn convert_ushort8_rte(double8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(double8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(double8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(double8);
-ushort8 __ovld __cnfn convert_ushort8_sat(double8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(double8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(double8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(double8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(double8);
-ushort16 __ovld __cnfn convert_ushort16(double16);
-ushort16 __ovld __cnfn convert_ushort16_rte(double16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(double16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(double16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(double16);
-ushort16 __ovld __cnfn convert_ushort16_sat(double16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(double16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(double16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(double16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(double16);
-
-int __ovld __cnfn convert_int(double);
-int __ovld __cnfn convert_int_rte(double);
-int __ovld __cnfn convert_int_rtn(double);
-int __ovld __cnfn convert_int_rtp(double);
-int __ovld __cnfn convert_int_rtz(double);
-int __ovld __cnfn convert_int_sat(double);
-int __ovld __cnfn convert_int_sat_rte(double);
-int __ovld __cnfn convert_int_sat_rtn(double);
-int __ovld __cnfn convert_int_sat_rtp(double);
-int __ovld __cnfn convert_int_sat_rtz(double);
-int2 __ovld __cnfn convert_int2(double2);
-int2 __ovld __cnfn convert_int2_rte(double2);
-int2 __ovld __cnfn convert_int2_rtn(double2);
-int2 __ovld __cnfn convert_int2_rtp(double2);
-int2 __ovld __cnfn convert_int2_rtz(double2);
-int2 __ovld __cnfn convert_int2_sat(double2);
-int2 __ovld __cnfn convert_int2_sat_rte(double2);
-int2 __ovld __cnfn convert_int2_sat_rtn(double2);
-int2 __ovld __cnfn convert_int2_sat_rtp(double2);
-int2 __ovld __cnfn convert_int2_sat_rtz(double2);
-int3 __ovld __cnfn convert_int3(double3);
-int3 __ovld __cnfn convert_int3_rte(double3);
-int3 __ovld __cnfn convert_int3_rtn(double3);
-int3 __ovld __cnfn convert_int3_rtp(double3);
-int3 __ovld __cnfn convert_int3_rtz(double3);
-int3 __ovld __cnfn convert_int3_sat(double3);
-int3 __ovld __cnfn convert_int3_sat_rte(double3);
-int3 __ovld __cnfn convert_int3_sat_rtn(double3);
-int3 __ovld __cnfn convert_int3_sat_rtp(double3);
-int3 __ovld __cnfn convert_int3_sat_rtz(double3);
-int4 __ovld __cnfn convert_int4(double4);
-int4 __ovld __cnfn convert_int4_rte(double4);
-int4 __ovld __cnfn convert_int4_rtn(double4);
-int4 __ovld __cnfn convert_int4_rtp(double4);
-int4 __ovld __cnfn convert_int4_rtz(double4);
-int4 __ovld __cnfn convert_int4_sat(double4);
-int4 __ovld __cnfn convert_int4_sat_rte(double4);
-int4 __ovld __cnfn convert_int4_sat_rtn(double4);
-int4 __ovld __cnfn convert_int4_sat_rtp(double4);
-int4 __ovld __cnfn convert_int4_sat_rtz(double4);
-int8 __ovld __cnfn convert_int8(double8);
-int8 __ovld __cnfn convert_int8_rte(double8);
-int8 __ovld __cnfn convert_int8_rtn(double8);
-int8 __ovld __cnfn convert_int8_rtp(double8);
-int8 __ovld __cnfn convert_int8_rtz(double8);
-int8 __ovld __cnfn convert_int8_sat(double8);
-int8 __ovld __cnfn convert_int8_sat_rte(double8);
-int8 __ovld __cnfn convert_int8_sat_rtn(double8);
-int8 __ovld __cnfn convert_int8_sat_rtp(double8);
-int8 __ovld __cnfn convert_int8_sat_rtz(double8);
-int16 __ovld __cnfn convert_int16(double16);
-int16 __ovld __cnfn convert_int16_rte(double16);
-int16 __ovld __cnfn convert_int16_rtn(double16);
-int16 __ovld __cnfn convert_int16_rtp(double16);
-int16 __ovld __cnfn convert_int16_rtz(double16);
-int16 __ovld __cnfn convert_int16_sat(double16);
-int16 __ovld __cnfn convert_int16_sat_rte(double16);
-int16 __ovld __cnfn convert_int16_sat_rtn(double16);
-int16 __ovld __cnfn convert_int16_sat_rtp(double16);
-int16 __ovld __cnfn convert_int16_sat_rtz(double16);
-
-uint __ovld __cnfn convert_uint(double);
-uint __ovld __cnfn convert_uint_rte(double);
-uint __ovld __cnfn convert_uint_rtn(double);
-uint __ovld __cnfn convert_uint_rtp(double);
-uint __ovld __cnfn convert_uint_rtz(double);
-uint __ovld __cnfn convert_uint_sat(double);
-uint __ovld __cnfn convert_uint_sat_rte(double);
-uint __ovld __cnfn convert_uint_sat_rtn(double);
-uint __ovld __cnfn convert_uint_sat_rtp(double);
-uint __ovld __cnfn convert_uint_sat_rtz(double);
-uint2 __ovld __cnfn convert_uint2(double2);
-uint2 __ovld __cnfn convert_uint2_rte(double2);
-uint2 __ovld __cnfn convert_uint2_rtn(double2);
-uint2 __ovld __cnfn convert_uint2_rtp(double2);
-uint2 __ovld __cnfn convert_uint2_rtz(double2);
-uint2 __ovld __cnfn convert_uint2_sat(double2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(double2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(double2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(double2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(double2);
-uint3 __ovld __cnfn convert_uint3(double3);
-uint3 __ovld __cnfn convert_uint3_rte(double3);
-uint3 __ovld __cnfn convert_uint3_rtn(double3);
-uint3 __ovld __cnfn convert_uint3_rtp(double3);
-uint3 __ovld __cnfn convert_uint3_rtz(double3);
-uint3 __ovld __cnfn convert_uint3_sat(double3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(double3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(double3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(double3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(double3);
-uint4 __ovld __cnfn convert_uint4(double4);
-uint4 __ovld __cnfn convert_uint4_rte(double4);
-uint4 __ovld __cnfn convert_uint4_rtn(double4);
-uint4 __ovld __cnfn convert_uint4_rtp(double4);
-uint4 __ovld __cnfn convert_uint4_rtz(double4);
-uint4 __ovld __cnfn convert_uint4_sat(double4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(double4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(double4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(double4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(double4);
-uint8 __ovld __cnfn convert_uint8(double8);
-uint8 __ovld __cnfn convert_uint8_rte(double8);
-uint8 __ovld __cnfn convert_uint8_rtn(double8);
-uint8 __ovld __cnfn convert_uint8_rtp(double8);
-uint8 __ovld __cnfn convert_uint8_rtz(double8);
-uint8 __ovld __cnfn convert_uint8_sat(double8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(double8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(double8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(double8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(double8);
-uint16 __ovld __cnfn convert_uint16(double16);
-uint16 __ovld __cnfn convert_uint16_rte(double16);
-uint16 __ovld __cnfn convert_uint16_rtn(double16);
-uint16 __ovld __cnfn convert_uint16_rtp(double16);
-uint16 __ovld __cnfn convert_uint16_rtz(double16);
-uint16 __ovld __cnfn convert_uint16_sat(double16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(double16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(double16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(double16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(double16);
-
-long __ovld __cnfn convert_long(double);
-long __ovld __cnfn convert_long_rte(double);
-long __ovld __cnfn convert_long_rtn(double);
-long __ovld __cnfn convert_long_rtp(double);
-long __ovld __cnfn convert_long_rtz(double);
-long __ovld __cnfn convert_long_sat(double);
-long __ovld __cnfn convert_long_sat_rte(double);
-long __ovld __cnfn convert_long_sat_rtn(double);
-long __ovld __cnfn convert_long_sat_rtp(double);
-long __ovld __cnfn convert_long_sat_rtz(double);
-long2 __ovld __cnfn convert_long2(double2);
-long2 __ovld __cnfn convert_long2_rte(double2);
-long2 __ovld __cnfn convert_long2_rtn(double2);
-long2 __ovld __cnfn convert_long2_rtp(double2);
-long2 __ovld __cnfn convert_long2_rtz(double2);
-long2 __ovld __cnfn convert_long2_sat(double2);
-long2 __ovld __cnfn convert_long2_sat_rte(double2);
-long2 __ovld __cnfn convert_long2_sat_rtn(double2);
-long2 __ovld __cnfn convert_long2_sat_rtp(double2);
-long2 __ovld __cnfn convert_long2_sat_rtz(double2);
-long3 __ovld __cnfn convert_long3(double3);
-long3 __ovld __cnfn convert_long3_rte(double3);
-long3 __ovld __cnfn convert_long3_rtn(double3);
-long3 __ovld __cnfn convert_long3_rtp(double3);
-long3 __ovld __cnfn convert_long3_rtz(double3);
-long3 __ovld __cnfn convert_long3_sat(double3);
-long3 __ovld __cnfn convert_long3_sat_rte(double3);
-long3 __ovld __cnfn convert_long3_sat_rtn(double3);
-long3 __ovld __cnfn convert_long3_sat_rtp(double3);
-long3 __ovld __cnfn convert_long3_sat_rtz(double3);
-long4 __ovld __cnfn convert_long4(double4);
-long4 __ovld __cnfn convert_long4_rte(double4);
-long4 __ovld __cnfn convert_long4_rtn(double4);
-long4 __ovld __cnfn convert_long4_rtp(double4);
-long4 __ovld __cnfn convert_long4_rtz(double4);
-long4 __ovld __cnfn convert_long4_sat(double4);
-long4 __ovld __cnfn convert_long4_sat_rte(double4);
-long4 __ovld __cnfn convert_long4_sat_rtn(double4);
-long4 __ovld __cnfn convert_long4_sat_rtp(double4);
-long4 __ovld __cnfn convert_long4_sat_rtz(double4);
-long8 __ovld __cnfn convert_long8(double8);
-long8 __ovld __cnfn convert_long8_rte(double8);
-long8 __ovld __cnfn convert_long8_rtn(double8);
-long8 __ovld __cnfn convert_long8_rtp(double8);
-long8 __ovld __cnfn convert_long8_rtz(double8);
-long8 __ovld __cnfn convert_long8_sat(double8);
-long8 __ovld __cnfn convert_long8_sat_rte(double8);
-long8 __ovld __cnfn convert_long8_sat_rtn(double8);
-long8 __ovld __cnfn convert_long8_sat_rtp(double8);
-long8 __ovld __cnfn convert_long8_sat_rtz(double8);
-long16 __ovld __cnfn convert_long16(double16);
-long16 __ovld __cnfn convert_long16_rte(double16);
-long16 __ovld __cnfn convert_long16_rtn(double16);
-long16 __ovld __cnfn convert_long16_rtp(double16);
-long16 __ovld __cnfn convert_long16_rtz(double16);
-long16 __ovld __cnfn convert_long16_sat(double16);
-long16 __ovld __cnfn convert_long16_sat_rte(double16);
-long16 __ovld __cnfn convert_long16_sat_rtn(double16);
-long16 __ovld __cnfn convert_long16_sat_rtp(double16);
-long16 __ovld __cnfn convert_long16_sat_rtz(double16);
-
-ulong __ovld __cnfn convert_ulong(double);
-ulong __ovld __cnfn convert_ulong_rte(double);
-ulong __ovld __cnfn convert_ulong_rtn(double);
-ulong __ovld __cnfn convert_ulong_rtp(double);
-ulong __ovld __cnfn convert_ulong_rtz(double);
-ulong __ovld __cnfn convert_ulong_sat(double);
-ulong __ovld __cnfn convert_ulong_sat_rte(double);
-ulong __ovld __cnfn convert_ulong_sat_rtn(double);
-ulong __ovld __cnfn convert_ulong_sat_rtp(double);
-ulong __ovld __cnfn convert_ulong_sat_rtz(double);
-ulong2 __ovld __cnfn convert_ulong2(double2);
-ulong2 __ovld __cnfn convert_ulong2_rte(double2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(double2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(double2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(double2);
-ulong2 __ovld __cnfn convert_ulong2_sat(double2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(double2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(double2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(double2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(double2);
-ulong3 __ovld __cnfn convert_ulong3(double3);
-ulong3 __ovld __cnfn convert_ulong3_rte(double3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(double3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(double3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(double3);
-ulong3 __ovld __cnfn convert_ulong3_sat(double3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(double3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(double3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(double3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(double3);
-ulong4 __ovld __cnfn convert_ulong4(double4);
-ulong4 __ovld __cnfn convert_ulong4_rte(double4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(double4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(double4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(double4);
-ulong4 __ovld __cnfn convert_ulong4_sat(double4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(double4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(double4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(double4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(double4);
-ulong8 __ovld __cnfn convert_ulong8(double8);
-ulong8 __ovld __cnfn convert_ulong8_rte(double8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(double8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(double8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(double8);
-ulong8 __ovld __cnfn convert_ulong8_sat(double8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(double8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(double8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(double8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(double8);
-ulong16 __ovld __cnfn convert_ulong16(double16);
-ulong16 __ovld __cnfn convert_ulong16_rte(double16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(double16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(double16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(double16);
-ulong16 __ovld __cnfn convert_ulong16_sat(double16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(double16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(double16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(double16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(double16);
-
-float __ovld __cnfn convert_float(double);
-float __ovld __cnfn convert_float_rte(double);
-float __ovld __cnfn convert_float_rtn(double);
-float __ovld __cnfn convert_float_rtp(double);
-float __ovld __cnfn convert_float_rtz(double);
-float2 __ovld __cnfn convert_float2(double2);
-float2 __ovld __cnfn convert_float2_rte(double2);
-float2 __ovld __cnfn convert_float2_rtn(double2);
-float2 __ovld __cnfn convert_float2_rtp(double2);
-float2 __ovld __cnfn convert_float2_rtz(double2);
-float3 __ovld __cnfn convert_float3(double3);
-float3 __ovld __cnfn convert_float3_rte(double3);
-float3 __ovld __cnfn convert_float3_rtn(double3);
-float3 __ovld __cnfn convert_float3_rtp(double3);
-float3 __ovld __cnfn convert_float3_rtz(double3);
-float4 __ovld __cnfn convert_float4(double4);
-float4 __ovld __cnfn convert_float4_rte(double4);
-float4 __ovld __cnfn convert_float4_rtn(double4);
-float4 __ovld __cnfn convert_float4_rtp(double4);
-float4 __ovld __cnfn convert_float4_rtz(double4);
-float8 __ovld __cnfn convert_float8(double8);
-float8 __ovld __cnfn convert_float8_rte(double8);
-float8 __ovld __cnfn convert_float8_rtn(double8);
-float8 __ovld __cnfn convert_float8_rtp(double8);
-float8 __ovld __cnfn convert_float8_rtz(double8);
-float16 __ovld __cnfn convert_float16(double16);
-float16 __ovld __cnfn convert_float16_rte(double16);
-float16 __ovld __cnfn convert_float16_rtn(double16);
-float16 __ovld __cnfn convert_float16_rtp(double16);
-float16 __ovld __cnfn convert_float16_rtz(double16);
-
-double __ovld __cnfn convert_double(char);
-double __ovld __cnfn convert_double(double);
-double __ovld __cnfn convert_double(float);
-double __ovld __cnfn convert_double(int);
-double __ovld __cnfn convert_double(long);
-double __ovld __cnfn convert_double(short);
-double __ovld __cnfn convert_double(uchar);
-double __ovld __cnfn convert_double(uint);
-double __ovld __cnfn convert_double(ulong);
-double __ovld __cnfn convert_double(ushort);
-double __ovld __cnfn convert_double_rte(char);
-double __ovld __cnfn convert_double_rte(double);
-double __ovld __cnfn convert_double_rte(float);
-double __ovld __cnfn convert_double_rte(int);
-double __ovld __cnfn convert_double_rte(long);
-double __ovld __cnfn convert_double_rte(short);
-double __ovld __cnfn convert_double_rte(uchar);
-double __ovld __cnfn convert_double_rte(uint);
-double __ovld __cnfn convert_double_rte(ulong);
-double __ovld __cnfn convert_double_rte(ushort);
-double __ovld __cnfn convert_double_rtn(char);
-double __ovld __cnfn convert_double_rtn(double);
-double __ovld __cnfn convert_double_rtn(float);
-double __ovld __cnfn convert_double_rtn(int);
-double __ovld __cnfn convert_double_rtn(long);
-double __ovld __cnfn convert_double_rtn(short);
-double __ovld __cnfn convert_double_rtn(uchar);
-double __ovld __cnfn convert_double_rtn(uint);
-double __ovld __cnfn convert_double_rtn(ulong);
-double __ovld __cnfn convert_double_rtn(ushort);
-double __ovld __cnfn convert_double_rtp(char);
-double __ovld __cnfn convert_double_rtp(double);
-double __ovld __cnfn convert_double_rtp(float);
-double __ovld __cnfn convert_double_rtp(int);
-double __ovld __cnfn convert_double_rtp(long);
-double __ovld __cnfn convert_double_rtp(short);
-double __ovld __cnfn convert_double_rtp(uchar);
-double __ovld __cnfn convert_double_rtp(uint);
-double __ovld __cnfn convert_double_rtp(ulong);
-double __ovld __cnfn convert_double_rtp(ushort);
-double __ovld __cnfn convert_double_rtz(char);
-double __ovld __cnfn convert_double_rtz(double);
-double __ovld __cnfn convert_double_rtz(float);
-double __ovld __cnfn convert_double_rtz(int);
-double __ovld __cnfn convert_double_rtz(long);
-double __ovld __cnfn convert_double_rtz(short);
-double __ovld __cnfn convert_double_rtz(uchar);
-double __ovld __cnfn convert_double_rtz(uint);
-double __ovld __cnfn convert_double_rtz(ulong);
-double __ovld __cnfn convert_double_rtz(ushort);
-double2 __ovld __cnfn convert_double2(char2);
-double2 __ovld __cnfn convert_double2(double2);
-double2 __ovld __cnfn convert_double2(float2);
-double2 __ovld __cnfn convert_double2(int2);
-double2 __ovld __cnfn convert_double2(long2);
-double2 __ovld __cnfn convert_double2(short2);
-double2 __ovld __cnfn convert_double2(uchar2);
-double2 __ovld __cnfn convert_double2(uint2);
-double2 __ovld __cnfn convert_double2(ulong2);
-double2 __ovld __cnfn convert_double2(ushort2);
-double2 __ovld __cnfn convert_double2_rte(char2);
-double2 __ovld __cnfn convert_double2_rte(double2);
-double2 __ovld __cnfn convert_double2_rte(float2);
-double2 __ovld __cnfn convert_double2_rte(int2);
-double2 __ovld __cnfn convert_double2_rte(long2);
-double2 __ovld __cnfn convert_double2_rte(short2);
-double2 __ovld __cnfn convert_double2_rte(uchar2);
-double2 __ovld __cnfn convert_double2_rte(uint2);
-double2 __ovld __cnfn convert_double2_rte(ulong2);
-double2 __ovld __cnfn convert_double2_rte(ushort2);
-double2 __ovld __cnfn convert_double2_rtn(char2);
-double2 __ovld __cnfn convert_double2_rtn(double2);
-double2 __ovld __cnfn convert_double2_rtn(float2);
-double2 __ovld __cnfn convert_double2_rtn(int2);
-double2 __ovld __cnfn convert_double2_rtn(long2);
-double2 __ovld __cnfn convert_double2_rtn(short2);
-double2 __ovld __cnfn convert_double2_rtn(uchar2);
-double2 __ovld __cnfn convert_double2_rtn(uint2);
-double2 __ovld __cnfn convert_double2_rtn(ulong2);
-double2 __ovld __cnfn convert_double2_rtn(ushort2);
-double2 __ovld __cnfn convert_double2_rtp(char2);
-double2 __ovld __cnfn convert_double2_rtp(double2);
-double2 __ovld __cnfn convert_double2_rtp(float2);
-double2 __ovld __cnfn convert_double2_rtp(int2);
-double2 __ovld __cnfn convert_double2_rtp(long2);
-double2 __ovld __cnfn convert_double2_rtp(short2);
-double2 __ovld __cnfn convert_double2_rtp(uchar2);
-double2 __ovld __cnfn convert_double2_rtp(uint2);
-double2 __ovld __cnfn convert_double2_rtp(ulong2);
-double2 __ovld __cnfn convert_double2_rtp(ushort2);
-double2 __ovld __cnfn convert_double2_rtz(char2);
-double2 __ovld __cnfn convert_double2_rtz(double2);
-double2 __ovld __cnfn convert_double2_rtz(float2);
-double2 __ovld __cnfn convert_double2_rtz(int2);
-double2 __ovld __cnfn convert_double2_rtz(long2);
-double2 __ovld __cnfn convert_double2_rtz(short2);
-double2 __ovld __cnfn convert_double2_rtz(uchar2);
-double2 __ovld __cnfn convert_double2_rtz(uint2);
-double2 __ovld __cnfn convert_double2_rtz(ulong2);
-double2 __ovld __cnfn convert_double2_rtz(ushort2);
-double3 __ovld __cnfn convert_double3(char3);
-double3 __ovld __cnfn convert_double3(double3);
-double3 __ovld __cnfn convert_double3(float3);
-double3 __ovld __cnfn convert_double3(int3);
-double3 __ovld __cnfn convert_double3(long3);
-double3 __ovld __cnfn convert_double3(short3);
-double3 __ovld __cnfn convert_double3(uchar3);
-double3 __ovld __cnfn convert_double3(uint3);
-double3 __ovld __cnfn convert_double3(ulong3);
-double3 __ovld __cnfn convert_double3(ushort3);
-double3 __ovld __cnfn convert_double3_rte(char3);
-double3 __ovld __cnfn convert_double3_rte(double3);
-double3 __ovld __cnfn convert_double3_rte(float3);
-double3 __ovld __cnfn convert_double3_rte(int3);
-double3 __ovld __cnfn convert_double3_rte(long3);
-double3 __ovld __cnfn convert_double3_rte(short3);
-double3 __ovld __cnfn convert_double3_rte(uchar3);
-double3 __ovld __cnfn convert_double3_rte(uint3);
-double3 __ovld __cnfn convert_double3_rte(ulong3);
-double3 __ovld __cnfn convert_double3_rte(ushort3);
-double3 __ovld __cnfn convert_double3_rtn(char3);
-double3 __ovld __cnfn convert_double3_rtn(double3);
-double3 __ovld __cnfn convert_double3_rtn(float3);
-double3 __ovld __cnfn convert_double3_rtn(int3);
-double3 __ovld __cnfn convert_double3_rtn(long3);
-double3 __ovld __cnfn convert_double3_rtn(short3);
-double3 __ovld __cnfn convert_double3_rtn(uchar3);
-double3 __ovld __cnfn convert_double3_rtn(uint3);
-double3 __ovld __cnfn convert_double3_rtn(ulong3);
-double3 __ovld __cnfn convert_double3_rtn(ushort3);
-double3 __ovld __cnfn convert_double3_rtp(char3);
-double3 __ovld __cnfn convert_double3_rtp(double3);
-double3 __ovld __cnfn convert_double3_rtp(float3);
-double3 __ovld __cnfn convert_double3_rtp(int3);
-double3 __ovld __cnfn convert_double3_rtp(long3);
-double3 __ovld __cnfn convert_double3_rtp(short3);
-double3 __ovld __cnfn convert_double3_rtp(uchar3);
-double3 __ovld __cnfn convert_double3_rtp(uint3);
-double3 __ovld __cnfn convert_double3_rtp(ulong3);
-double3 __ovld __cnfn convert_double3_rtp(ushort3);
-double3 __ovld __cnfn convert_double3_rtz(char3);
-double3 __ovld __cnfn convert_double3_rtz(double3);
-double3 __ovld __cnfn convert_double3_rtz(float3);
-double3 __ovld __cnfn convert_double3_rtz(int3);
-double3 __ovld __cnfn convert_double3_rtz(long3);
-double3 __ovld __cnfn convert_double3_rtz(short3);
-double3 __ovld __cnfn convert_double3_rtz(uchar3);
-double3 __ovld __cnfn convert_double3_rtz(uint3);
-double3 __ovld __cnfn convert_double3_rtz(ulong3);
-double3 __ovld __cnfn convert_double3_rtz(ushort3);
-double4 __ovld __cnfn convert_double4(char4);
-double4 __ovld __cnfn convert_double4(double4);
-double4 __ovld __cnfn convert_double4(float4);
-double4 __ovld __cnfn convert_double4(int4);
-double4 __ovld __cnfn convert_double4(long4);
-double4 __ovld __cnfn convert_double4(short4);
-double4 __ovld __cnfn convert_double4(uchar4);
-double4 __ovld __cnfn convert_double4(uint4);
-double4 __ovld __cnfn convert_double4(ulong4);
-double4 __ovld __cnfn convert_double4(ushort4);
-double4 __ovld __cnfn convert_double4_rte(char4);
-double4 __ovld __cnfn convert_double4_rte(double4);
-double4 __ovld __cnfn convert_double4_rte(float4);
-double4 __ovld __cnfn convert_double4_rte(int4);
-double4 __ovld __cnfn convert_double4_rte(long4);
-double4 __ovld __cnfn convert_double4_rte(short4);
-double4 __ovld __cnfn convert_double4_rte(uchar4);
-double4 __ovld __cnfn convert_double4_rte(uint4);
-double4 __ovld __cnfn convert_double4_rte(ulong4);
-double4 __ovld __cnfn convert_double4_rte(ushort4);
-double4 __ovld __cnfn convert_double4_rtn(char4);
-double4 __ovld __cnfn convert_double4_rtn(double4);
-double4 __ovld __cnfn convert_double4_rtn(float4);
-double4 __ovld __cnfn convert_double4_rtn(int4);
-double4 __ovld __cnfn convert_double4_rtn(long4);
-double4 __ovld __cnfn convert_double4_rtn(short4);
-double4 __ovld __cnfn convert_double4_rtn(uchar4);
-double4 __ovld __cnfn convert_double4_rtn(uint4);
-double4 __ovld __cnfn convert_double4_rtn(ulong4);
-double4 __ovld __cnfn convert_double4_rtn(ushort4);
-double4 __ovld __cnfn convert_double4_rtp(char4);
-double4 __ovld __cnfn convert_double4_rtp(double4);
-double4 __ovld __cnfn convert_double4_rtp(float4);
-double4 __ovld __cnfn convert_double4_rtp(int4);
-double4 __ovld __cnfn convert_double4_rtp(long4);
-double4 __ovld __cnfn convert_double4_rtp(short4);
-double4 __ovld __cnfn convert_double4_rtp(uchar4);
-double4 __ovld __cnfn convert_double4_rtp(uint4);
-double4 __ovld __cnfn convert_double4_rtp(ulong4);
-double4 __ovld __cnfn convert_double4_rtp(ushort4);
-double4 __ovld __cnfn convert_double4_rtz(char4);
-double4 __ovld __cnfn convert_double4_rtz(double4);
-double4 __ovld __cnfn convert_double4_rtz(float4);
-double4 __ovld __cnfn convert_double4_rtz(int4);
-double4 __ovld __cnfn convert_double4_rtz(long4);
-double4 __ovld __cnfn convert_double4_rtz(short4);
-double4 __ovld __cnfn convert_double4_rtz(uchar4);
-double4 __ovld __cnfn convert_double4_rtz(uint4);
-double4 __ovld __cnfn convert_double4_rtz(ulong4);
-double4 __ovld __cnfn convert_double4_rtz(ushort4);
-double8 __ovld __cnfn convert_double8(char8);
-double8 __ovld __cnfn convert_double8(double8);
-double8 __ovld __cnfn convert_double8(float8);
-double8 __ovld __cnfn convert_double8(int8);
-double8 __ovld __cnfn convert_double8(long8);
-double8 __ovld __cnfn convert_double8(short8);
-double8 __ovld __cnfn convert_double8(uchar8);
-double8 __ovld __cnfn convert_double8(uint8);
-double8 __ovld __cnfn convert_double8(ulong8);
-double8 __ovld __cnfn convert_double8(ushort8);
-double8 __ovld __cnfn convert_double8_rte(char8);
-double8 __ovld __cnfn convert_double8_rte(double8);
-double8 __ovld __cnfn convert_double8_rte(float8);
-double8 __ovld __cnfn convert_double8_rte(int8);
-double8 __ovld __cnfn convert_double8_rte(long8);
-double8 __ovld __cnfn convert_double8_rte(short8);
-double8 __ovld __cnfn convert_double8_rte(uchar8);
-double8 __ovld __cnfn convert_double8_rte(uint8);
-double8 __ovld __cnfn convert_double8_rte(ulong8);
-double8 __ovld __cnfn convert_double8_rte(ushort8);
-double8 __ovld __cnfn convert_double8_rtn(char8);
-double8 __ovld __cnfn convert_double8_rtn(double8);
-double8 __ovld __cnfn convert_double8_rtn(float8);
-double8 __ovld __cnfn convert_double8_rtn(int8);
-double8 __ovld __cnfn convert_double8_rtn(long8);
-double8 __ovld __cnfn convert_double8_rtn(short8);
-double8 __ovld __cnfn convert_double8_rtn(uchar8);
-double8 __ovld __cnfn convert_double8_rtn(uint8);
-double8 __ovld __cnfn convert_double8_rtn(ulong8);
-double8 __ovld __cnfn convert_double8_rtn(ushort8);
-double8 __ovld __cnfn convert_double8_rtp(char8);
-double8 __ovld __cnfn convert_double8_rtp(double8);
-double8 __ovld __cnfn convert_double8_rtp(float8);
-double8 __ovld __cnfn convert_double8_rtp(int8);
-double8 __ovld __cnfn convert_double8_rtp(long8);
-double8 __ovld __cnfn convert_double8_rtp(short8);
-double8 __ovld __cnfn convert_double8_rtp(uchar8);
-double8 __ovld __cnfn convert_double8_rtp(uint8);
-double8 __ovld __cnfn convert_double8_rtp(ulong8);
-double8 __ovld __cnfn convert_double8_rtp(ushort8);
-double8 __ovld __cnfn convert_double8_rtz(char8);
-double8 __ovld __cnfn convert_double8_rtz(double8);
-double8 __ovld __cnfn convert_double8_rtz(float8);
-double8 __ovld __cnfn convert_double8_rtz(int8);
-double8 __ovld __cnfn convert_double8_rtz(long8);
-double8 __ovld __cnfn convert_double8_rtz(short8);
-double8 __ovld __cnfn convert_double8_rtz(uchar8);
-double8 __ovld __cnfn convert_double8_rtz(uint8);
-double8 __ovld __cnfn convert_double8_rtz(ulong8);
-double8 __ovld __cnfn convert_double8_rtz(ushort8);
-double16 __ovld __cnfn convert_double16(char16);
-double16 __ovld __cnfn convert_double16(double16);
-double16 __ovld __cnfn convert_double16(float16);
-double16 __ovld __cnfn convert_double16(int16);
-double16 __ovld __cnfn convert_double16(long16);
-double16 __ovld __cnfn convert_double16(short16);
-double16 __ovld __cnfn convert_double16(uchar16);
-double16 __ovld __cnfn convert_double16(uint16);
-double16 __ovld __cnfn convert_double16(ulong16);
-double16 __ovld __cnfn convert_double16(ushort16);
-double16 __ovld __cnfn convert_double16_rte(char16);
-double16 __ovld __cnfn convert_double16_rte(double16);
-double16 __ovld __cnfn convert_double16_rte(float16);
-double16 __ovld __cnfn convert_double16_rte(int16);
-double16 __ovld __cnfn convert_double16_rte(long16);
-double16 __ovld __cnfn convert_double16_rte(short16);
-double16 __ovld __cnfn convert_double16_rte(uchar16);
-double16 __ovld __cnfn convert_double16_rte(uint16);
-double16 __ovld __cnfn convert_double16_rte(ulong16);
-double16 __ovld __cnfn convert_double16_rte(ushort16);
-double16 __ovld __cnfn convert_double16_rtn(char16);
-double16 __ovld __cnfn convert_double16_rtn(double16);
-double16 __ovld __cnfn convert_double16_rtn(float16);
-double16 __ovld __cnfn convert_double16_rtn(int16);
-double16 __ovld __cnfn convert_double16_rtn(long16);
-double16 __ovld __cnfn convert_double16_rtn(short16);
-double16 __ovld __cnfn convert_double16_rtn(uchar16);
-double16 __ovld __cnfn convert_double16_rtn(uint16);
-double16 __ovld __cnfn convert_double16_rtn(ulong16);
-double16 __ovld __cnfn convert_double16_rtn(ushort16);
-double16 __ovld __cnfn convert_double16_rtp(char16);
-double16 __ovld __cnfn convert_double16_rtp(double16);
-double16 __ovld __cnfn convert_double16_rtp(float16);
-double16 __ovld __cnfn convert_double16_rtp(int16);
-double16 __ovld __cnfn convert_double16_rtp(long16);
-double16 __ovld __cnfn convert_double16_rtp(short16);
-double16 __ovld __cnfn convert_double16_rtp(uchar16);
-double16 __ovld __cnfn convert_double16_rtp(uint16);
-double16 __ovld __cnfn convert_double16_rtp(ulong16);
-double16 __ovld __cnfn convert_double16_rtp(ushort16);
-double16 __ovld __cnfn convert_double16_rtz(char16);
-double16 __ovld __cnfn convert_double16_rtz(double16);
-double16 __ovld __cnfn convert_double16_rtz(float16);
-double16 __ovld __cnfn convert_double16_rtz(int16);
-double16 __ovld __cnfn convert_double16_rtz(long16);
-double16 __ovld __cnfn convert_double16_rtz(short16);
-double16 __ovld __cnfn convert_double16_rtz(uchar16);
-double16 __ovld __cnfn convert_double16_rtz(uint16);
-double16 __ovld __cnfn convert_double16_rtz(ulong16);
-double16 __ovld __cnfn convert_double16_rtz(ushort16);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-// Convert half types to non-double types.
-uchar __ovld __cnfn convert_uchar(half);
-uchar __ovld __cnfn convert_uchar_rte(half);
-uchar __ovld __cnfn convert_uchar_rtp(half);
-uchar __ovld __cnfn convert_uchar_rtn(half);
-uchar __ovld __cnfn convert_uchar_rtz(half);
-uchar __ovld __cnfn convert_uchar_sat(half);
-uchar __ovld __cnfn convert_uchar_sat_rte(half);
-uchar __ovld __cnfn convert_uchar_sat_rtp(half);
-uchar __ovld __cnfn convert_uchar_sat_rtn(half);
-uchar __ovld __cnfn convert_uchar_sat_rtz(half);
-uchar2 __ovld __cnfn convert_uchar2(half2);
-uchar2 __ovld __cnfn convert_uchar2_rte(half2);
-uchar2 __ovld __cnfn convert_uchar2_rtp(half2);
-uchar2 __ovld __cnfn convert_uchar2_rtn(half2);
-uchar2 __ovld __cnfn convert_uchar2_rtz(half2);
-uchar2 __ovld __cnfn convert_uchar2_sat(half2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rte(half2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtp(half2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtn(half2);
-uchar2 __ovld __cnfn convert_uchar2_sat_rtz(half2);
-uchar3 __ovld __cnfn convert_uchar3(half3);
-uchar3 __ovld __cnfn convert_uchar3_rte(half3);
-uchar3 __ovld __cnfn convert_uchar3_rtp(half3);
-uchar3 __ovld __cnfn convert_uchar3_rtn(half3);
-uchar3 __ovld __cnfn convert_uchar3_rtz(half3);
-uchar3 __ovld __cnfn convert_uchar3_sat(half3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rte(half3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtp(half3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtn(half3);
-uchar3 __ovld __cnfn convert_uchar3_sat_rtz(half3);
-uchar4 __ovld __cnfn convert_uchar4(half4);
-uchar4 __ovld __cnfn convert_uchar4_rte(half4);
-uchar4 __ovld __cnfn convert_uchar4_rtp(half4);
-uchar4 __ovld __cnfn convert_uchar4_rtn(half4);
-uchar4 __ovld __cnfn convert_uchar4_rtz(half4);
-uchar4 __ovld __cnfn convert_uchar4_sat(half4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rte(half4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtp(half4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtn(half4);
-uchar4 __ovld __cnfn convert_uchar4_sat_rtz(half4);
-uchar8 __ovld __cnfn convert_uchar8(half8);
-uchar8 __ovld __cnfn convert_uchar8_rte(half8);
-uchar8 __ovld __cnfn convert_uchar8_rtp(half8);
-uchar8 __ovld __cnfn convert_uchar8_rtn(half8);
-uchar8 __ovld __cnfn convert_uchar8_rtz(half8);
-uchar8 __ovld __cnfn convert_uchar8_sat(half8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rte(half8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtp(half8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtn(half8);
-uchar8 __ovld __cnfn convert_uchar8_sat_rtz(half8);
-uchar16 __ovld __cnfn convert_uchar16(half16);
-uchar16 __ovld __cnfn convert_uchar16_rte(half16);
-uchar16 __ovld __cnfn convert_uchar16_rtp(half16);
-uchar16 __ovld __cnfn convert_uchar16_rtn(half16);
-uchar16 __ovld __cnfn convert_uchar16_rtz(half16);
-uchar16 __ovld __cnfn convert_uchar16_sat(half16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rte(half16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtp(half16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtn(half16);
-uchar16 __ovld __cnfn convert_uchar16_sat_rtz(half16);
-ushort __ovld __cnfn convert_ushort(half);
-ushort __ovld __cnfn convert_ushort_rte(half);
-ushort __ovld __cnfn convert_ushort_rtp(half);
-ushort __ovld __cnfn convert_ushort_rtn(half);
-ushort __ovld __cnfn convert_ushort_rtz(half);
-ushort __ovld __cnfn convert_ushort_sat(half);
-ushort __ovld __cnfn convert_ushort_sat_rte(half);
-ushort __ovld __cnfn convert_ushort_sat_rtp(half);
-ushort __ovld __cnfn convert_ushort_sat_rtn(half);
-ushort __ovld __cnfn convert_ushort_sat_rtz(half);
-ushort2 __ovld __cnfn convert_ushort2(half2);
-ushort2 __ovld __cnfn convert_ushort2_rte(half2);
-ushort2 __ovld __cnfn convert_ushort2_rtp(half2);
-ushort2 __ovld __cnfn convert_ushort2_rtn(half2);
-ushort2 __ovld __cnfn convert_ushort2_rtz(half2);
-ushort2 __ovld __cnfn convert_ushort2_sat(half2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rte(half2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtp(half2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtn(half2);
-ushort2 __ovld __cnfn convert_ushort2_sat_rtz(half2);
-ushort3 __ovld __cnfn convert_ushort3(half3);
-ushort3 __ovld __cnfn convert_ushort3_rte(half3);
-ushort3 __ovld __cnfn convert_ushort3_rtp(half3);
-ushort3 __ovld __cnfn convert_ushort3_rtn(half3);
-ushort3 __ovld __cnfn convert_ushort3_rtz(half3);
-ushort3 __ovld __cnfn convert_ushort3_sat(half3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rte(half3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtp(half3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtn(half3);
-ushort3 __ovld __cnfn convert_ushort3_sat_rtz(half3);
-ushort4 __ovld __cnfn convert_ushort4(half4);
-ushort4 __ovld __cnfn convert_ushort4_rte(half4);
-ushort4 __ovld __cnfn convert_ushort4_rtp(half4);
-ushort4 __ovld __cnfn convert_ushort4_rtn(half4);
-ushort4 __ovld __cnfn convert_ushort4_rtz(half4);
-ushort4 __ovld __cnfn convert_ushort4_sat(half4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rte(half4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtp(half4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtn(half4);
-ushort4 __ovld __cnfn convert_ushort4_sat_rtz(half4);
-ushort8 __ovld __cnfn convert_ushort8(half8);
-ushort8 __ovld __cnfn convert_ushort8_rte(half8);
-ushort8 __ovld __cnfn convert_ushort8_rtp(half8);
-ushort8 __ovld __cnfn convert_ushort8_rtn(half8);
-ushort8 __ovld __cnfn convert_ushort8_rtz(half8);
-ushort8 __ovld __cnfn convert_ushort8_sat(half8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rte(half8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtp(half8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtn(half8);
-ushort8 __ovld __cnfn convert_ushort8_sat_rtz(half8);
-ushort16 __ovld __cnfn convert_ushort16(half16);
-ushort16 __ovld __cnfn convert_ushort16_rte(half16);
-ushort16 __ovld __cnfn convert_ushort16_rtp(half16);
-ushort16 __ovld __cnfn convert_ushort16_rtn(half16);
-ushort16 __ovld __cnfn convert_ushort16_rtz(half16);
-ushort16 __ovld __cnfn convert_ushort16_sat(half16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rte(half16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtp(half16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtn(half16);
-ushort16 __ovld __cnfn convert_ushort16_sat_rtz(half16);
-uint __ovld __cnfn convert_uint(half);
-uint __ovld __cnfn convert_uint_rte(half);
-uint __ovld __cnfn convert_uint_rtp(half);
-uint __ovld __cnfn convert_uint_rtn(half);
-uint __ovld __cnfn convert_uint_rtz(half);
-uint __ovld __cnfn convert_uint_sat(half);
-uint __ovld __cnfn convert_uint_sat_rte(half);
-uint __ovld __cnfn convert_uint_sat_rtp(half);
-uint __ovld __cnfn convert_uint_sat_rtn(half);
-uint __ovld __cnfn convert_uint_sat_rtz(half);
-uint2 __ovld __cnfn convert_uint2(half2);
-uint2 __ovld __cnfn convert_uint2_rte(half2);
-uint2 __ovld __cnfn convert_uint2_rtp(half2);
-uint2 __ovld __cnfn convert_uint2_rtn(half2);
-uint2 __ovld __cnfn convert_uint2_rtz(half2);
-uint2 __ovld __cnfn convert_uint2_sat(half2);
-uint2 __ovld __cnfn convert_uint2_sat_rte(half2);
-uint2 __ovld __cnfn convert_uint2_sat_rtp(half2);
-uint2 __ovld __cnfn convert_uint2_sat_rtn(half2);
-uint2 __ovld __cnfn convert_uint2_sat_rtz(half2);
-uint3 __ovld __cnfn convert_uint3(half3);
-uint3 __ovld __cnfn convert_uint3_rte(half3);
-uint3 __ovld __cnfn convert_uint3_rtp(half3);
-uint3 __ovld __cnfn convert_uint3_rtn(half3);
-uint3 __ovld __cnfn convert_uint3_rtz(half3);
-uint3 __ovld __cnfn convert_uint3_sat(half3);
-uint3 __ovld __cnfn convert_uint3_sat_rte(half3);
-uint3 __ovld __cnfn convert_uint3_sat_rtp(half3);
-uint3 __ovld __cnfn convert_uint3_sat_rtn(half3);
-uint3 __ovld __cnfn convert_uint3_sat_rtz(half3);
-uint4 __ovld __cnfn convert_uint4(half4);
-uint4 __ovld __cnfn convert_uint4_rte(half4);
-uint4 __ovld __cnfn convert_uint4_rtp(half4);
-uint4 __ovld __cnfn convert_uint4_rtn(half4);
-uint4 __ovld __cnfn convert_uint4_rtz(half4);
-uint4 __ovld __cnfn convert_uint4_sat(half4);
-uint4 __ovld __cnfn convert_uint4_sat_rte(half4);
-uint4 __ovld __cnfn convert_uint4_sat_rtp(half4);
-uint4 __ovld __cnfn convert_uint4_sat_rtn(half4);
-uint4 __ovld __cnfn convert_uint4_sat_rtz(half4);
-uint8 __ovld __cnfn convert_uint8(half8);
-uint8 __ovld __cnfn convert_uint8_rte(half8);
-uint8 __ovld __cnfn convert_uint8_rtp(half8);
-uint8 __ovld __cnfn convert_uint8_rtn(half8);
-uint8 __ovld __cnfn convert_uint8_rtz(half8);
-uint8 __ovld __cnfn convert_uint8_sat(half8);
-uint8 __ovld __cnfn convert_uint8_sat_rte(half8);
-uint8 __ovld __cnfn convert_uint8_sat_rtp(half8);
-uint8 __ovld __cnfn convert_uint8_sat_rtn(half8);
-uint8 __ovld __cnfn convert_uint8_sat_rtz(half8);
-uint16 __ovld __cnfn convert_uint16(half16);
-uint16 __ovld __cnfn convert_uint16_rte(half16);
-uint16 __ovld __cnfn convert_uint16_rtp(half16);
-uint16 __ovld __cnfn convert_uint16_rtn(half16);
-uint16 __ovld __cnfn convert_uint16_rtz(half16);
-uint16 __ovld __cnfn convert_uint16_sat(half16);
-uint16 __ovld __cnfn convert_uint16_sat_rte(half16);
-uint16 __ovld __cnfn convert_uint16_sat_rtp(half16);
-uint16 __ovld __cnfn convert_uint16_sat_rtn(half16);
-uint16 __ovld __cnfn convert_uint16_sat_rtz(half16);
-ulong __ovld __cnfn convert_ulong(half);
-ulong __ovld __cnfn convert_ulong_rte(half);
-ulong __ovld __cnfn convert_ulong_rtp(half);
-ulong __ovld __cnfn convert_ulong_rtn(half);
-ulong __ovld __cnfn convert_ulong_rtz(half);
-ulong __ovld __cnfn convert_ulong_sat(half);
-ulong __ovld __cnfn convert_ulong_sat_rte(half);
-ulong __ovld __cnfn convert_ulong_sat_rtp(half);
-ulong __ovld __cnfn convert_ulong_sat_rtn(half);
-ulong __ovld __cnfn convert_ulong_sat_rtz(half);
-ulong2 __ovld __cnfn convert_ulong2(half2);
-ulong2 __ovld __cnfn convert_ulong2_rte(half2);
-ulong2 __ovld __cnfn convert_ulong2_rtp(half2);
-ulong2 __ovld __cnfn convert_ulong2_rtn(half2);
-ulong2 __ovld __cnfn convert_ulong2_rtz(half2);
-ulong2 __ovld __cnfn convert_ulong2_sat(half2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rte(half2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtp(half2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtn(half2);
-ulong2 __ovld __cnfn convert_ulong2_sat_rtz(half2);
-ulong3 __ovld __cnfn convert_ulong3(half3);
-ulong3 __ovld __cnfn convert_ulong3_rte(half3);
-ulong3 __ovld __cnfn convert_ulong3_rtp(half3);
-ulong3 __ovld __cnfn convert_ulong3_rtn(half3);
-ulong3 __ovld __cnfn convert_ulong3_rtz(half3);
-ulong3 __ovld __cnfn convert_ulong3_sat(half3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rte(half3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtp(half3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtn(half3);
-ulong3 __ovld __cnfn convert_ulong3_sat_rtz(half3);
-ulong4 __ovld __cnfn convert_ulong4(half4);
-ulong4 __ovld __cnfn convert_ulong4_rte(half4);
-ulong4 __ovld __cnfn convert_ulong4_rtp(half4);
-ulong4 __ovld __cnfn convert_ulong4_rtn(half4);
-ulong4 __ovld __cnfn convert_ulong4_rtz(half4);
-ulong4 __ovld __cnfn convert_ulong4_sat(half4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rte(half4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtp(half4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtn(half4);
-ulong4 __ovld __cnfn convert_ulong4_sat_rtz(half4);
-ulong8 __ovld __cnfn convert_ulong8(half8);
-ulong8 __ovld __cnfn convert_ulong8_rte(half8);
-ulong8 __ovld __cnfn convert_ulong8_rtp(half8);
-ulong8 __ovld __cnfn convert_ulong8_rtn(half8);
-ulong8 __ovld __cnfn convert_ulong8_rtz(half8);
-ulong8 __ovld __cnfn convert_ulong8_sat(half8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rte(half8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtp(half8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtn(half8);
-ulong8 __ovld __cnfn convert_ulong8_sat_rtz(half8);
-ulong16 __ovld __cnfn convert_ulong16(half16);
-ulong16 __ovld __cnfn convert_ulong16_rte(half16);
-ulong16 __ovld __cnfn convert_ulong16_rtp(half16);
-ulong16 __ovld __cnfn convert_ulong16_rtn(half16);
-ulong16 __ovld __cnfn convert_ulong16_rtz(half16);
-ulong16 __ovld __cnfn convert_ulong16_sat(half16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rte(half16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtp(half16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtn(half16);
-ulong16 __ovld __cnfn convert_ulong16_sat_rtz(half16);
-char __ovld __cnfn convert_char(half);
-char __ovld __cnfn convert_char_rte(half);
-char __ovld __cnfn convert_char_rtp(half);
-char __ovld __cnfn convert_char_rtn(half);
-char __ovld __cnfn convert_char_rtz(half);
-char __ovld __cnfn convert_char_sat(half);
-char __ovld __cnfn convert_char_sat_rte(half);
-char __ovld __cnfn convert_char_sat_rtp(half);
-char __ovld __cnfn convert_char_sat_rtn(half);
-char __ovld __cnfn convert_char_sat_rtz(half);
-char2 __ovld __cnfn convert_char2(half2);
-char2 __ovld __cnfn convert_char2_rte(half2);
-char2 __ovld __cnfn convert_char2_rtp(half2);
-char2 __ovld __cnfn convert_char2_rtn(half2);
-char2 __ovld __cnfn convert_char2_rtz(half2);
-char2 __ovld __cnfn convert_char2_sat(half2);
-char2 __ovld __cnfn convert_char2_sat_rte(half2);
-char2 __ovld __cnfn convert_char2_sat_rtp(half2);
-char2 __ovld __cnfn convert_char2_sat_rtn(half2);
-char2 __ovld __cnfn convert_char2_sat_rtz(half2);
-char3 __ovld __cnfn convert_char3(half3);
-char3 __ovld __cnfn convert_char3_rte(half3);
-char3 __ovld __cnfn convert_char3_rtp(half3);
-char3 __ovld __cnfn convert_char3_rtn(half3);
-char3 __ovld __cnfn convert_char3_rtz(half3);
-char3 __ovld __cnfn convert_char3_sat(half3);
-char3 __ovld __cnfn convert_char3_sat_rte(half3);
-char3 __ovld __cnfn convert_char3_sat_rtp(half3);
-char3 __ovld __cnfn convert_char3_sat_rtn(half3);
-char3 __ovld __cnfn convert_char3_sat_rtz(half3);
-char4 __ovld __cnfn convert_char4(half4);
-char4 __ovld __cnfn convert_char4_rte(half4);
-char4 __ovld __cnfn convert_char4_rtp(half4);
-char4 __ovld __cnfn convert_char4_rtn(half4);
-char4 __ovld __cnfn convert_char4_rtz(half4);
-char4 __ovld __cnfn convert_char4_sat(half4);
-char4 __ovld __cnfn convert_char4_sat_rte(half4);
-char4 __ovld __cnfn convert_char4_sat_rtp(half4);
-char4 __ovld __cnfn convert_char4_sat_rtn(half4);
-char4 __ovld __cnfn convert_char4_sat_rtz(half4);
-char8 __ovld __cnfn convert_char8(half8);
-char8 __ovld __cnfn convert_char8_rte(half8);
-char8 __ovld __cnfn convert_char8_rtp(half8);
-char8 __ovld __cnfn convert_char8_rtn(half8);
-char8 __ovld __cnfn convert_char8_rtz(half8);
-char8 __ovld __cnfn convert_char8_sat(half8);
-char8 __ovld __cnfn convert_char8_sat_rte(half8);
-char8 __ovld __cnfn convert_char8_sat_rtp(half8);
-char8 __ovld __cnfn convert_char8_sat_rtn(half8);
-char8 __ovld __cnfn convert_char8_sat_rtz(half8);
-char16 __ovld __cnfn convert_char16(half16);
-char16 __ovld __cnfn convert_char16_rte(half16);
-char16 __ovld __cnfn convert_char16_rtp(half16);
-char16 __ovld __cnfn convert_char16_rtn(half16);
-char16 __ovld __cnfn convert_char16_rtz(half16);
-char16 __ovld __cnfn convert_char16_sat(half16);
-char16 __ovld __cnfn convert_char16_sat_rte(half16);
-char16 __ovld __cnfn convert_char16_sat_rtp(half16);
-char16 __ovld __cnfn convert_char16_sat_rtn(half16);
-char16 __ovld __cnfn convert_char16_sat_rtz(half16);
-short __ovld __cnfn convert_short(half);
-short __ovld __cnfn convert_short_rte(half);
-short __ovld __cnfn convert_short_rtp(half);
-short __ovld __cnfn convert_short_rtn(half);
-short __ovld __cnfn convert_short_rtz(half);
-short __ovld __cnfn convert_short_sat(half);
-short __ovld __cnfn convert_short_sat_rte(half);
-short __ovld __cnfn convert_short_sat_rtp(half);
-short __ovld __cnfn convert_short_sat_rtn(half);
-short __ovld __cnfn convert_short_sat_rtz(half);
-short2 __ovld __cnfn convert_short2(half2);
-short2 __ovld __cnfn convert_short2_rte(half2);
-short2 __ovld __cnfn convert_short2_rtp(half2);
-short2 __ovld __cnfn convert_short2_rtn(half2);
-short2 __ovld __cnfn convert_short2_rtz(half2);
-short2 __ovld __cnfn convert_short2_sat(half2);
-short2 __ovld __cnfn convert_short2_sat_rte(half2);
-short2 __ovld __cnfn convert_short2_sat_rtp(half2);
-short2 __ovld __cnfn convert_short2_sat_rtn(half2);
-short2 __ovld __cnfn convert_short2_sat_rtz(half2);
-short3 __ovld __cnfn convert_short3(half3);
-short3 __ovld __cnfn convert_short3_rte(half3);
-short3 __ovld __cnfn convert_short3_rtp(half3);
-short3 __ovld __cnfn convert_short3_rtn(half3);
-short3 __ovld __cnfn convert_short3_rtz(half3);
-short3 __ovld __cnfn convert_short3_sat(half3);
-short3 __ovld __cnfn convert_short3_sat_rte(half3);
-short3 __ovld __cnfn convert_short3_sat_rtp(half3);
-short3 __ovld __cnfn convert_short3_sat_rtn(half3);
-short3 __ovld __cnfn convert_short3_sat_rtz(half3);
-short4 __ovld __cnfn convert_short4(half4);
-short4 __ovld __cnfn convert_short4_rte(half4);
-short4 __ovld __cnfn convert_short4_rtp(half4);
-short4 __ovld __cnfn convert_short4_rtn(half4);
-short4 __ovld __cnfn convert_short4_rtz(half4);
-short4 __ovld __cnfn convert_short4_sat(half4);
-short4 __ovld __cnfn convert_short4_sat_rte(half4);
-short4 __ovld __cnfn convert_short4_sat_rtp(half4);
-short4 __ovld __cnfn convert_short4_sat_rtn(half4);
-short4 __ovld __cnfn convert_short4_sat_rtz(half4);
-short8 __ovld __cnfn convert_short8(half8);
-short8 __ovld __cnfn convert_short8_rte(half8);
-short8 __ovld __cnfn convert_short8_rtp(half8);
-short8 __ovld __cnfn convert_short8_rtn(half8);
-short8 __ovld __cnfn convert_short8_rtz(half8);
-short8 __ovld __cnfn convert_short8_sat(half8);
-short8 __ovld __cnfn convert_short8_sat_rte(half8);
-short8 __ovld __cnfn convert_short8_sat_rtp(half8);
-short8 __ovld __cnfn convert_short8_sat_rtn(half8);
-short8 __ovld __cnfn convert_short8_sat_rtz(half8);
-short16 __ovld __cnfn convert_short16(half16);
-short16 __ovld __cnfn convert_short16_rte(half16);
-short16 __ovld __cnfn convert_short16_rtp(half16);
-short16 __ovld __cnfn convert_short16_rtn(half16);
-short16 __ovld __cnfn convert_short16_rtz(half16);
-short16 __ovld __cnfn convert_short16_sat(half16);
-short16 __ovld __cnfn convert_short16_sat_rte(half16);
-short16 __ovld __cnfn convert_short16_sat_rtp(half16);
-short16 __ovld __cnfn convert_short16_sat_rtn(half16);
-short16 __ovld __cnfn convert_short16_sat_rtz(half16);
-int __ovld __cnfn convert_int(half);
-int __ovld __cnfn convert_int_rte(half);
-int __ovld __cnfn convert_int_rtp(half);
-int __ovld __cnfn convert_int_rtn(half);
-int __ovld __cnfn convert_int_rtz(half);
-int __ovld __cnfn convert_int_sat(half);
-int __ovld __cnfn convert_int_sat_rte(half);
-int __ovld __cnfn convert_int_sat_rtp(half);
-int __ovld __cnfn convert_int_sat_rtn(half);
-int __ovld __cnfn convert_int_sat_rtz(half);
-int2 __ovld __cnfn convert_int2(half2);
-int2 __ovld __cnfn convert_int2_rte(half2);
-int2 __ovld __cnfn convert_int2_rtp(half2);
-int2 __ovld __cnfn convert_int2_rtn(half2);
-int2 __ovld __cnfn convert_int2_rtz(half2);
-int2 __ovld __cnfn convert_int2_sat(half2);
-int2 __ovld __cnfn convert_int2_sat_rte(half2);
-int2 __ovld __cnfn convert_int2_sat_rtp(half2);
-int2 __ovld __cnfn convert_int2_sat_rtn(half2);
-int2 __ovld __cnfn convert_int2_sat_rtz(half2);
-int3 __ovld __cnfn convert_int3(half3);
-int3 __ovld __cnfn convert_int3_rte(half3);
-int3 __ovld __cnfn convert_int3_rtp(half3);
-int3 __ovld __cnfn convert_int3_rtn(half3);
-int3 __ovld __cnfn convert_int3_rtz(half3);
-int3 __ovld __cnfn convert_int3_sat(half3);
-int3 __ovld __cnfn convert_int3_sat_rte(half3);
-int3 __ovld __cnfn convert_int3_sat_rtp(half3);
-int3 __ovld __cnfn convert_int3_sat_rtn(half3);
-int3 __ovld __cnfn convert_int3_sat_rtz(half3);
-int4 __ovld __cnfn convert_int4(half4);
-int4 __ovld __cnfn convert_int4_rte(half4);
-int4 __ovld __cnfn convert_int4_rtp(half4);
-int4 __ovld __cnfn convert_int4_rtn(half4);
-int4 __ovld __cnfn convert_int4_rtz(half4);
-int4 __ovld __cnfn convert_int4_sat(half4);
-int4 __ovld __cnfn convert_int4_sat_rte(half4);
-int4 __ovld __cnfn convert_int4_sat_rtp(half4);
-int4 __ovld __cnfn convert_int4_sat_rtn(half4);
-int4 __ovld __cnfn convert_int4_sat_rtz(half4);
-int8 __ovld __cnfn convert_int8(half8);
-int8 __ovld __cnfn convert_int8_rte(half8);
-int8 __ovld __cnfn convert_int8_rtp(half8);
-int8 __ovld __cnfn convert_int8_rtn(half8);
-int8 __ovld __cnfn convert_int8_rtz(half8);
-int8 __ovld __cnfn convert_int8_sat(half8);
-int8 __ovld __cnfn convert_int8_sat_rte(half8);
-int8 __ovld __cnfn convert_int8_sat_rtp(half8);
-int8 __ovld __cnfn convert_int8_sat_rtn(half8);
-int8 __ovld __cnfn convert_int8_sat_rtz(half8);
-int16 __ovld __cnfn convert_int16(half16);
-int16 __ovld __cnfn convert_int16_rte(half16);
-int16 __ovld __cnfn convert_int16_rtp(half16);
-int16 __ovld __cnfn convert_int16_rtn(half16);
-int16 __ovld __cnfn convert_int16_rtz(half16);
-int16 __ovld __cnfn convert_int16_sat(half16);
-int16 __ovld __cnfn convert_int16_sat_rte(half16);
-int16 __ovld __cnfn convert_int16_sat_rtp(half16);
-int16 __ovld __cnfn convert_int16_sat_rtn(half16);
-int16 __ovld __cnfn convert_int16_sat_rtz(half16);
-long __ovld __cnfn convert_long(half);
-long __ovld __cnfn convert_long_rte(half);
-long __ovld __cnfn convert_long_rtp(half);
-long __ovld __cnfn convert_long_rtn(half);
-long __ovld __cnfn convert_long_rtz(half);
-long __ovld __cnfn convert_long_sat(half);
-long __ovld __cnfn convert_long_sat_rte(half);
-long __ovld __cnfn convert_long_sat_rtp(half);
-long __ovld __cnfn convert_long_sat_rtn(half);
-long __ovld __cnfn convert_long_sat_rtz(half);
-long2 __ovld __cnfn convert_long2(half2);
-long2 __ovld __cnfn convert_long2_rte(half2);
-long2 __ovld __cnfn convert_long2_rtp(half2);
-long2 __ovld __cnfn convert_long2_rtn(half2);
-long2 __ovld __cnfn convert_long2_rtz(half2);
-long2 __ovld __cnfn convert_long2_sat(half2);
-long2 __ovld __cnfn convert_long2_sat_rte(half2);
-long2 __ovld __cnfn convert_long2_sat_rtp(half2);
-long2 __ovld __cnfn convert_long2_sat_rtn(half2);
-long2 __ovld __cnfn convert_long2_sat_rtz(half2);
-long3 __ovld __cnfn convert_long3(half3);
-long3 __ovld __cnfn convert_long3_rte(half3);
-long3 __ovld __cnfn convert_long3_rtp(half3);
-long3 __ovld __cnfn convert_long3_rtn(half3);
-long3 __ovld __cnfn convert_long3_rtz(half3);
-long3 __ovld __cnfn convert_long3_sat(half3);
-long3 __ovld __cnfn convert_long3_sat_rte(half3);
-long3 __ovld __cnfn convert_long3_sat_rtp(half3);
-long3 __ovld __cnfn convert_long3_sat_rtn(half3);
-long3 __ovld __cnfn convert_long3_sat_rtz(half3);
-long4 __ovld __cnfn convert_long4(half4);
-long4 __ovld __cnfn convert_long4_rte(half4);
-long4 __ovld __cnfn convert_long4_rtp(half4);
-long4 __ovld __cnfn convert_long4_rtn(half4);
-long4 __ovld __cnfn convert_long4_rtz(half4);
-long4 __ovld __cnfn convert_long4_sat(half4);
-long4 __ovld __cnfn convert_long4_sat_rte(half4);
-long4 __ovld __cnfn convert_long4_sat_rtp(half4);
-long4 __ovld __cnfn convert_long4_sat_rtn(half4);
-long4 __ovld __cnfn convert_long4_sat_rtz(half4);
-long8 __ovld __cnfn convert_long8(half8);
-long8 __ovld __cnfn convert_long8_rte(half8);
-long8 __ovld __cnfn convert_long8_rtp(half8);
-long8 __ovld __cnfn convert_long8_rtn(half8);
-long8 __ovld __cnfn convert_long8_rtz(half8);
-long8 __ovld __cnfn convert_long8_sat(half8);
-long8 __ovld __cnfn convert_long8_sat_rte(half8);
-long8 __ovld __cnfn convert_long8_sat_rtp(half8);
-long8 __ovld __cnfn convert_long8_sat_rtn(half8);
-long8 __ovld __cnfn convert_long8_sat_rtz(half8);
-long16 __ovld __cnfn convert_long16(half16);
-long16 __ovld __cnfn convert_long16_rte(half16);
-long16 __ovld __cnfn convert_long16_rtp(half16);
-long16 __ovld __cnfn convert_long16_rtn(half16);
-long16 __ovld __cnfn convert_long16_rtz(half16);
-long16 __ovld __cnfn convert_long16_sat(half16);
-long16 __ovld __cnfn convert_long16_sat_rte(half16);
-long16 __ovld __cnfn convert_long16_sat_rtp(half16);
-long16 __ovld __cnfn convert_long16_sat_rtn(half16);
-long16 __ovld __cnfn convert_long16_sat_rtz(half16);
-float __ovld __cnfn convert_float(half);
-float __ovld __cnfn convert_float_rte(half);
-float __ovld __cnfn convert_float_rtp(half);
-float __ovld __cnfn convert_float_rtn(half);
-float __ovld __cnfn convert_float_rtz(half);
-float2 __ovld __cnfn convert_float2(half2);
-float2 __ovld __cnfn convert_float2_rte(half2);
-float2 __ovld __cnfn convert_float2_rtp(half2);
-float2 __ovld __cnfn convert_float2_rtn(half2);
-float2 __ovld __cnfn convert_float2_rtz(half2);
-float3 __ovld __cnfn convert_float3(half3);
-float3 __ovld __cnfn convert_float3_rte(half3);
-float3 __ovld __cnfn convert_float3_rtp(half3);
-float3 __ovld __cnfn convert_float3_rtn(half3);
-float3 __ovld __cnfn convert_float3_rtz(half3);
-float4 __ovld __cnfn convert_float4(half4);
-float4 __ovld __cnfn convert_float4_rte(half4);
-float4 __ovld __cnfn convert_float4_rtp(half4);
-float4 __ovld __cnfn convert_float4_rtn(half4);
-float4 __ovld __cnfn convert_float4_rtz(half4);
-float8 __ovld __cnfn convert_float8(half8);
-float8 __ovld __cnfn convert_float8_rte(half8);
-float8 __ovld __cnfn convert_float8_rtp(half8);
-float8 __ovld __cnfn convert_float8_rtn(half8);
-float8 __ovld __cnfn convert_float8_rtz(half8);
-float16 __ovld __cnfn convert_float16(half16);
-float16 __ovld __cnfn convert_float16_rte(half16);
-float16 __ovld __cnfn convert_float16_rtp(half16);
-float16 __ovld __cnfn convert_float16_rtn(half16);
-float16 __ovld __cnfn convert_float16_rtz(half16);
-
-// Convert non-double types to half types.
-half __ovld __cnfn convert_half(uchar);
-half __ovld __cnfn convert_half(ushort);
-half __ovld __cnfn convert_half(uint);
-half __ovld __cnfn convert_half(ulong);
-half __ovld __cnfn convert_half(char);
-half __ovld __cnfn convert_half(short);
-half __ovld __cnfn convert_half(int);
-half __ovld __cnfn convert_half(long);
-half __ovld __cnfn convert_half(float);
-half __ovld __cnfn convert_half(half);
-half __ovld __cnfn convert_half_rte(uchar);
-half __ovld __cnfn convert_half_rte(ushort);
-half __ovld __cnfn convert_half_rte(uint);
-half __ovld __cnfn convert_half_rte(ulong);
-half __ovld __cnfn convert_half_rte(char);
-half __ovld __cnfn convert_half_rte(short);
-half __ovld __cnfn convert_half_rte(int);
-half __ovld __cnfn convert_half_rte(long);
-half __ovld __cnfn convert_half_rte(float);
-half __ovld __cnfn convert_half_rte(half);
-half __ovld __cnfn convert_half_rtp(uchar);
-half __ovld __cnfn convert_half_rtp(ushort);
-half __ovld __cnfn convert_half_rtp(uint);
-half __ovld __cnfn convert_half_rtp(ulong);
-half __ovld __cnfn convert_half_rtp(char);
-half __ovld __cnfn convert_half_rtp(short);
-half __ovld __cnfn convert_half_rtp(int);
-half __ovld __cnfn convert_half_rtp(long);
-half __ovld __cnfn convert_half_rtp(float);
-half __ovld __cnfn convert_half_rtp(half);
-half __ovld __cnfn convert_half_rtn(uchar);
-half __ovld __cnfn convert_half_rtn(ushort);
-half __ovld __cnfn convert_half_rtn(uint);
-half __ovld __cnfn convert_half_rtn(ulong);
-half __ovld __cnfn convert_half_rtn(char);
-half __ovld __cnfn convert_half_rtn(short);
-half __ovld __cnfn convert_half_rtn(int);
-half __ovld __cnfn convert_half_rtn(long);
-half __ovld __cnfn convert_half_rtn(float);
-half __ovld __cnfn convert_half_rtn(half);
-half __ovld __cnfn convert_half_rtz(uchar);
-half __ovld __cnfn convert_half_rtz(ushort);
-half __ovld __cnfn convert_half_rtz(uint);
-half __ovld __cnfn convert_half_rtz(ulong);
-half __ovld __cnfn convert_half_rtz(char);
-half __ovld __cnfn convert_half_rtz(short);
-half __ovld __cnfn convert_half_rtz(int);
-half __ovld __cnfn convert_half_rtz(long);
-half __ovld __cnfn convert_half_rtz(float);
-half __ovld __cnfn convert_half_rtz(half);
-half2 __ovld __cnfn convert_half2(char2);
-half2 __ovld __cnfn convert_half2(uchar2);
-half2 __ovld __cnfn convert_half2(short2);
-half2 __ovld __cnfn convert_half2(ushort2);
-half2 __ovld __cnfn convert_half2(int2);
-half2 __ovld __cnfn convert_half2(uint2);
-half2 __ovld __cnfn convert_half2(long2);
-half2 __ovld __cnfn convert_half2(ulong2);
-half2 __ovld __cnfn convert_half2(float2);
-half2 __ovld __cnfn convert_half2(half2);
-half2 __ovld __cnfn convert_half2_rte(char2);
-half2 __ovld __cnfn convert_half2_rte(uchar2);
-half2 __ovld __cnfn convert_half2_rte(short2);
-half2 __ovld __cnfn convert_half2_rte(ushort2);
-half2 __ovld __cnfn convert_half2_rte(int2);
-half2 __ovld __cnfn convert_half2_rte(uint2);
-half2 __ovld __cnfn convert_half2_rte(long2);
-half2 __ovld __cnfn convert_half2_rte(ulong2);
-half2 __ovld __cnfn convert_half2_rte(float2);
-half2 __ovld __cnfn convert_half2_rte(half2);
-half2 __ovld __cnfn convert_half2_rtp(char2);
-half2 __ovld __cnfn convert_half2_rtp(uchar2);
-half2 __ovld __cnfn convert_half2_rtp(short2);
-half2 __ovld __cnfn convert_half2_rtp(ushort2);
-half2 __ovld __cnfn convert_half2_rtp(int2);
-half2 __ovld __cnfn convert_half2_rtp(uint2);
-half2 __ovld __cnfn convert_half2_rtp(long2);
-half2 __ovld __cnfn convert_half2_rtp(ulong2);
-half2 __ovld __cnfn convert_half2_rtp(float2);
-half2 __ovld __cnfn convert_half2_rtp(half2);
-half2 __ovld __cnfn convert_half2_rtn(char2);
-half2 __ovld __cnfn convert_half2_rtn(uchar2);
-half2 __ovld __cnfn convert_half2_rtn(short2);
-half2 __ovld __cnfn convert_half2_rtn(ushort2);
-half2 __ovld __cnfn convert_half2_rtn(int2);
-half2 __ovld __cnfn convert_half2_rtn(uint2);
-half2 __ovld __cnfn convert_half2_rtn(long2);
-half2 __ovld __cnfn convert_half2_rtn(ulong2);
-half2 __ovld __cnfn convert_half2_rtn(float2);
-half2 __ovld __cnfn convert_half2_rtn(half2);
-half2 __ovld __cnfn convert_half2_rtz(char2);
-half2 __ovld __cnfn convert_half2_rtz(uchar2);
-half2 __ovld __cnfn convert_half2_rtz(short2);
-half2 __ovld __cnfn convert_half2_rtz(ushort2);
-half2 __ovld __cnfn convert_half2_rtz(int2);
-half2 __ovld __cnfn convert_half2_rtz(uint2);
-half2 __ovld __cnfn convert_half2_rtz(long2);
-half2 __ovld __cnfn convert_half2_rtz(ulong2);
-half2 __ovld __cnfn convert_half2_rtz(float2);
-half2 __ovld __cnfn convert_half2_rtz(half2);
-half3 __ovld __cnfn convert_half3(char3);
-half3 __ovld __cnfn convert_half3(uchar3);
-half3 __ovld __cnfn convert_half3(short3);
-half3 __ovld __cnfn convert_half3(ushort3);
-half3 __ovld __cnfn convert_half3(int3);
-half3 __ovld __cnfn convert_half3(uint3);
-half3 __ovld __cnfn convert_half3(long3);
-half3 __ovld __cnfn convert_half3(ulong3);
-half3 __ovld __cnfn convert_half3(float3);
-half3 __ovld __cnfn convert_half3(half3);
-half3 __ovld __cnfn convert_half3_rte(char3);
-half3 __ovld __cnfn convert_half3_rte(uchar3);
-half3 __ovld __cnfn convert_half3_rte(short3);
-half3 __ovld __cnfn convert_half3_rte(ushort3);
-half3 __ovld __cnfn convert_half3_rte(int3);
-half3 __ovld __cnfn convert_half3_rte(uint3);
-half3 __ovld __cnfn convert_half3_rte(long3);
-half3 __ovld __cnfn convert_half3_rte(ulong3);
-half3 __ovld __cnfn convert_half3_rte(float3);
-half3 __ovld __cnfn convert_half3_rte(half3);
-half3 __ovld __cnfn convert_half3_rtp(char3);
-half3 __ovld __cnfn convert_half3_rtp(uchar3);
-half3 __ovld __cnfn convert_half3_rtp(short3);
-half3 __ovld __cnfn convert_half3_rtp(ushort3);
-half3 __ovld __cnfn convert_half3_rtp(int3);
-half3 __ovld __cnfn convert_half3_rtp(uint3);
-half3 __ovld __cnfn convert_half3_rtp(long3);
-half3 __ovld __cnfn convert_half3_rtp(ulong3);
-half3 __ovld __cnfn convert_half3_rtp(float3);
-half3 __ovld __cnfn convert_half3_rtp(half3);
-half3 __ovld __cnfn convert_half3_rtn(char3);
-half3 __ovld __cnfn convert_half3_rtn(uchar3);
-half3 __ovld __cnfn convert_half3_rtn(short3);
-half3 __ovld __cnfn convert_half3_rtn(ushort3);
-half3 __ovld __cnfn convert_half3_rtn(int3);
-half3 __ovld __cnfn convert_half3_rtn(uint3);
-half3 __ovld __cnfn convert_half3_rtn(long3);
-half3 __ovld __cnfn convert_half3_rtn(ulong3);
-half3 __ovld __cnfn convert_half3_rtn(float3);
-half3 __ovld __cnfn convert_half3_rtn(half3);
-half3 __ovld __cnfn convert_half3_rtz(char3);
-half3 __ovld __cnfn convert_half3_rtz(uchar3);
-half3 __ovld __cnfn convert_half3_rtz(short3);
-half3 __ovld __cnfn convert_half3_rtz(ushort3);
-half3 __ovld __cnfn convert_half3_rtz(int3);
-half3 __ovld __cnfn convert_half3_rtz(uint3);
-half3 __ovld __cnfn convert_half3_rtz(long3);
-half3 __ovld __cnfn convert_half3_rtz(ulong3);
-half3 __ovld __cnfn convert_half3_rtz(float3);
-half3 __ovld __cnfn convert_half3_rtz(half3);
-half4 __ovld __cnfn convert_half4(char4);
-half4 __ovld __cnfn convert_half4(uchar4);
-half4 __ovld __cnfn convert_half4(short4);
-half4 __ovld __cnfn convert_half4(ushort4);
-half4 __ovld __cnfn convert_half4(int4);
-half4 __ovld __cnfn convert_half4(uint4);
-half4 __ovld __cnfn convert_half4(long4);
-half4 __ovld __cnfn convert_half4(ulong4);
-half4 __ovld __cnfn convert_half4(float4);
-half4 __ovld __cnfn convert_half4(half4);
-half4 __ovld __cnfn convert_half4_rte(char4);
-half4 __ovld __cnfn convert_half4_rte(uchar4);
-half4 __ovld __cnfn convert_half4_rte(short4);
-half4 __ovld __cnfn convert_half4_rte(ushort4);
-half4 __ovld __cnfn convert_half4_rte(int4);
-half4 __ovld __cnfn convert_half4_rte(uint4);
-half4 __ovld __cnfn convert_half4_rte(long4);
-half4 __ovld __cnfn convert_half4_rte(ulong4);
-half4 __ovld __cnfn convert_half4_rte(float4);
-half4 __ovld __cnfn convert_half4_rte(half4);
-half4 __ovld __cnfn convert_half4_rtp(char4);
-half4 __ovld __cnfn convert_half4_rtp(uchar4);
-half4 __ovld __cnfn convert_half4_rtp(short4);
-half4 __ovld __cnfn convert_half4_rtp(ushort4);
-half4 __ovld __cnfn convert_half4_rtp(int4);
-half4 __ovld __cnfn convert_half4_rtp(uint4);
-half4 __ovld __cnfn convert_half4_rtp(long4);
-half4 __ovld __cnfn convert_half4_rtp(ulong4);
-half4 __ovld __cnfn convert_half4_rtp(float4);
-half4 __ovld __cnfn convert_half4_rtp(half4);
-half4 __ovld __cnfn convert_half4_rtn(char4);
-half4 __ovld __cnfn convert_half4_rtn(uchar4);
-half4 __ovld __cnfn convert_half4_rtn(short4);
-half4 __ovld __cnfn convert_half4_rtn(ushort4);
-half4 __ovld __cnfn convert_half4_rtn(int4);
-half4 __ovld __cnfn convert_half4_rtn(uint4);
-half4 __ovld __cnfn convert_half4_rtn(long4);
-half4 __ovld __cnfn convert_half4_rtn(ulong4);
-half4 __ovld __cnfn convert_half4_rtn(float4);
-half4 __ovld __cnfn convert_half4_rtn(half4);
-half4 __ovld __cnfn convert_half4_rtz(char4);
-half4 __ovld __cnfn convert_half4_rtz(uchar4);
-half4 __ovld __cnfn convert_half4_rtz(short4);
-half4 __ovld __cnfn convert_half4_rtz(ushort4);
-half4 __ovld __cnfn convert_half4_rtz(int4);
-half4 __ovld __cnfn convert_half4_rtz(uint4);
-half4 __ovld __cnfn convert_half4_rtz(long4);
-half4 __ovld __cnfn convert_half4_rtz(ulong4);
-half4 __ovld __cnfn convert_half4_rtz(float4);
-half4 __ovld __cnfn convert_half4_rtz(half4);
-half8 __ovld __cnfn convert_half8(char8);
-half8 __ovld __cnfn convert_half8(uchar8);
-half8 __ovld __cnfn convert_half8(short8);
-half8 __ovld __cnfn convert_half8(ushort8);
-half8 __ovld __cnfn convert_half8(int8);
-half8 __ovld __cnfn convert_half8(uint8);
-half8 __ovld __cnfn convert_half8(long8);
-half8 __ovld __cnfn convert_half8(ulong8);
-half8 __ovld __cnfn convert_half8(float8);
-half8 __ovld __cnfn convert_half8(half8);
-half8 __ovld __cnfn convert_half8_rte(char8);
-half8 __ovld __cnfn convert_half8_rte(uchar8);
-half8 __ovld __cnfn convert_half8_rte(short8);
-half8 __ovld __cnfn convert_half8_rte(ushort8);
-half8 __ovld __cnfn convert_half8_rte(int8);
-half8 __ovld __cnfn convert_half8_rte(uint8);
-half8 __ovld __cnfn convert_half8_rte(long8);
-half8 __ovld __cnfn convert_half8_rte(ulong8);
-half8 __ovld __cnfn convert_half8_rte(float8);
-half8 __ovld __cnfn convert_half8_rte(half8);
-half8 __ovld __cnfn convert_half8_rtp(char8);
-half8 __ovld __cnfn convert_half8_rtp(uchar8);
-half8 __ovld __cnfn convert_half8_rtp(short8);
-half8 __ovld __cnfn convert_half8_rtp(ushort8);
-half8 __ovld __cnfn convert_half8_rtp(int8);
-half8 __ovld __cnfn convert_half8_rtp(uint8);
-half8 __ovld __cnfn convert_half8_rtp(long8);
-half8 __ovld __cnfn convert_half8_rtp(ulong8);
-half8 __ovld __cnfn convert_half8_rtp(float8);
-half8 __ovld __cnfn convert_half8_rtp(half8);
-half8 __ovld __cnfn convert_half8_rtn(char8);
-half8 __ovld __cnfn convert_half8_rtn(uchar8);
-half8 __ovld __cnfn convert_half8_rtn(short8);
-half8 __ovld __cnfn convert_half8_rtn(ushort8);
-half8 __ovld __cnfn convert_half8_rtn(int8);
-half8 __ovld __cnfn convert_half8_rtn(uint8);
-half8 __ovld __cnfn convert_half8_rtn(long8);
-half8 __ovld __cnfn convert_half8_rtn(ulong8);
-half8 __ovld __cnfn convert_half8_rtn(float8);
-half8 __ovld __cnfn convert_half8_rtn(half8);
-half8 __ovld __cnfn convert_half8_rtz(char8);
-half8 __ovld __cnfn convert_half8_rtz(uchar8);
-half8 __ovld __cnfn convert_half8_rtz(short8);
-half8 __ovld __cnfn convert_half8_rtz(ushort8);
-half8 __ovld __cnfn convert_half8_rtz(int8);
-half8 __ovld __cnfn convert_half8_rtz(uint8);
-half8 __ovld __cnfn convert_half8_rtz(long8);
-half8 __ovld __cnfn convert_half8_rtz(ulong8);
-half8 __ovld __cnfn convert_half8_rtz(float8);
-half8 __ovld __cnfn convert_half8_rtz(half8);
-half16 __ovld __cnfn convert_half16(char16);
-half16 __ovld __cnfn convert_half16(uchar16);
-half16 __ovld __cnfn convert_half16(short16);
-half16 __ovld __cnfn convert_half16(ushort16);
-half16 __ovld __cnfn convert_half16(int16);
-half16 __ovld __cnfn convert_half16(uint16);
-half16 __ovld __cnfn convert_half16(long16);
-half16 __ovld __cnfn convert_half16(ulong16);
-half16 __ovld __cnfn convert_half16(float16);
-half16 __ovld __cnfn convert_half16(half16);
-half16 __ovld __cnfn convert_half16_rte(char16);
-half16 __ovld __cnfn convert_half16_rte(uchar16);
-half16 __ovld __cnfn convert_half16_rte(short16);
-half16 __ovld __cnfn convert_half16_rte(ushort16);
-half16 __ovld __cnfn convert_half16_rte(int16);
-half16 __ovld __cnfn convert_half16_rte(uint16);
-half16 __ovld __cnfn convert_half16_rte(long16);
-half16 __ovld __cnfn convert_half16_rte(ulong16);
-half16 __ovld __cnfn convert_half16_rte(float16);
-half16 __ovld __cnfn convert_half16_rte(half16);
-half16 __ovld __cnfn convert_half16_rtp(char16);
-half16 __ovld __cnfn convert_half16_rtp(uchar16);
-half16 __ovld __cnfn convert_half16_rtp(short16);
-half16 __ovld __cnfn convert_half16_rtp(ushort16);
-half16 __ovld __cnfn convert_half16_rtp(int16);
-half16 __ovld __cnfn convert_half16_rtp(uint16);
-half16 __ovld __cnfn convert_half16_rtp(long16);
-half16 __ovld __cnfn convert_half16_rtp(ulong16);
-half16 __ovld __cnfn convert_half16_rtp(float16);
-half16 __ovld __cnfn convert_half16_rtp(half16);
-half16 __ovld __cnfn convert_half16_rtn(char16);
-half16 __ovld __cnfn convert_half16_rtn(uchar16);
-half16 __ovld __cnfn convert_half16_rtn(short16);
-half16 __ovld __cnfn convert_half16_rtn(ushort16);
-half16 __ovld __cnfn convert_half16_rtn(int16);
-half16 __ovld __cnfn convert_half16_rtn(uint16);
-half16 __ovld __cnfn convert_half16_rtn(long16);
-half16 __ovld __cnfn convert_half16_rtn(ulong16);
-half16 __ovld __cnfn convert_half16_rtn(float16);
-half16 __ovld __cnfn convert_half16_rtn(half16);
-half16 __ovld __cnfn convert_half16_rtz(char16);
-half16 __ovld __cnfn convert_half16_rtz(uchar16);
-half16 __ovld __cnfn convert_half16_rtz(short16);
-half16 __ovld __cnfn convert_half16_rtz(ushort16);
-half16 __ovld __cnfn convert_half16_rtz(int16);
-half16 __ovld __cnfn convert_half16_rtz(uint16);
-half16 __ovld __cnfn convert_half16_rtz(long16);
-half16 __ovld __cnfn convert_half16_rtz(ulong16);
-half16 __ovld __cnfn convert_half16_rtz(float16);
-half16 __ovld __cnfn convert_half16_rtz(half16);
-
-// Convert half types to double types.
-#ifdef cl_khr_fp64
-double __ovld __cnfn convert_double(half);
-double __ovld __cnfn convert_double_rte(half);
-double __ovld __cnfn convert_double_rtp(half);
-double __ovld __cnfn convert_double_rtn(half);
-double __ovld __cnfn convert_double_rtz(half);
-double2 __ovld __cnfn convert_double2(half2);
-double2 __ovld __cnfn convert_double2_rte(half2);
-double2 __ovld __cnfn convert_double2_rtp(half2);
-double2 __ovld __cnfn convert_double2_rtn(half2);
-double2 __ovld __cnfn convert_double2_rtz(half2);
-double3 __ovld __cnfn convert_double3(half3);
-double3 __ovld __cnfn convert_double3_rte(half3);
-double3 __ovld __cnfn convert_double3_rtp(half3);
-double3 __ovld __cnfn convert_double3_rtn(half3);
-double3 __ovld __cnfn convert_double3_rtz(half3);
-double4 __ovld __cnfn convert_double4(half4);
-double4 __ovld __cnfn convert_double4_rte(half4);
-double4 __ovld __cnfn convert_double4_rtp(half4);
-double4 __ovld __cnfn convert_double4_rtn(half4);
-double4 __ovld __cnfn convert_double4_rtz(half4);
-double8 __ovld __cnfn convert_double8(half8);
-double8 __ovld __cnfn convert_double8_rte(half8);
-double8 __ovld __cnfn convert_double8_rtp(half8);
-double8 __ovld __cnfn convert_double8_rtn(half8);
-double8 __ovld __cnfn convert_double8_rtz(half8);
-double16 __ovld __cnfn convert_double16(half16);
-double16 __ovld __cnfn convert_double16_rte(half16);
-double16 __ovld __cnfn convert_double16_rtp(half16);
-double16 __ovld __cnfn convert_double16_rtn(half16);
-double16 __ovld __cnfn convert_double16_rtz(half16);
-
-// Convert double types to half types.
-half __ovld __cnfn convert_half(double);
-half __ovld __cnfn convert_half_rte(double);
-half __ovld __cnfn convert_half_rtp(double);
-half __ovld __cnfn convert_half_rtn(double);
-half __ovld __cnfn convert_half_rtz(double);
-half2 __ovld __cnfn convert_half2(double2);
-half2 __ovld __cnfn convert_half2_rte(double2);
-half2 __ovld __cnfn convert_half2_rtp(double2);
-half2 __ovld __cnfn convert_half2_rtn(double2);
-half2 __ovld __cnfn convert_half2_rtz(double2);
-half3 __ovld __cnfn convert_half3(double3);
-half3 __ovld __cnfn convert_half3_rte(double3);
-half3 __ovld __cnfn convert_half3_rtp(double3);
-half3 __ovld __cnfn convert_half3_rtn(double3);
-half3 __ovld __cnfn convert_half3_rtz(double3);
-half4 __ovld __cnfn convert_half4(double4);
-half4 __ovld __cnfn convert_half4_rte(double4);
-half4 __ovld __cnfn convert_half4_rtp(double4);
-half4 __ovld __cnfn convert_half4_rtn(double4);
-half4 __ovld __cnfn convert_half4_rtz(double4);
-half8 __ovld __cnfn convert_half8(double8);
-half8 __ovld __cnfn convert_half8_rte(double8);
-half8 __ovld __cnfn convert_half8_rtp(double8);
-half8 __ovld __cnfn convert_half8_rtn(double8);
-half8 __ovld __cnfn convert_half8_rtz(double8);
-half16 __ovld __cnfn convert_half16(double16);
-half16 __ovld __cnfn convert_half16_rte(double16);
-half16 __ovld __cnfn convert_half16_rtp(double16);
-half16 __ovld __cnfn convert_half16_rtn(double16);
-half16 __ovld __cnfn convert_half16_rtz(double16);
-#endif //cl_khr_fp64
-
-#endif // cl_khr_fp16
-
-/**
- * OpenCL v1.1/1.2/2.0 s6.2.4.2 - as_type operators
- * Reinterprets a data type as another data type of the same size
- */
-#define as_char(x) __builtin_astype((x),   char)
-#define as_char2(x) __builtin_astype((x),  char2)
-#define as_char3(x) __builtin_astype((x),  char3)
-#define as_char4(x) __builtin_astype((x),  char4)
-#define as_char8(x) __builtin_astype((x),  char8)
-#define as_char16(x) __builtin_astype((x), char16)
-
-#define as_uchar(x) __builtin_astype((x),   uchar)
-#define as_uchar2(x) __builtin_astype((x),  uchar2)
-#define as_uchar3(x) __builtin_astype((x),  uchar3)
-#define as_uchar4(x) __builtin_astype((x),  uchar4)
-#define as_uchar8(x) __builtin_astype((x),  uchar8)
-#define as_uchar16(x) __builtin_astype((x), uchar16)
-
-#define as_short(x) __builtin_astype((x),   short)
-#define as_short2(x) __builtin_astype((x),  short2)
-#define as_short3(x) __builtin_astype((x),  short3)
-#define as_short4(x) __builtin_astype((x),  short4)
-#define as_short8(x) __builtin_astype((x),  short8)
-#define as_short16(x) __builtin_astype((x), short16)
-
-#define as_ushort(x) __builtin_astype((x),   ushort)
-#define as_ushort2(x) __builtin_astype((x),  ushort2)
-#define as_ushort3(x) __builtin_astype((x),  ushort3)
-#define as_ushort4(x) __builtin_astype((x),  ushort4)
-#define as_ushort8(x) __builtin_astype((x),  ushort8)
-#define as_ushort16(x) __builtin_astype((x), ushort16)
-
-#define as_int(x) __builtin_astype((x),   int)
-#define as_int2(x) __builtin_astype((x),  int2)
-#define as_int3(x) __builtin_astype((x),  int3)
-#define as_int4(x) __builtin_astype((x),  int4)
-#define as_int8(x) __builtin_astype((x),  int8)
-#define as_int16(x) __builtin_astype((x), int16)
-
-#define as_uint(x) __builtin_astype((x),   uint)
-#define as_uint2(x) __builtin_astype((x),  uint2)
-#define as_uint3(x) __builtin_astype((x),  uint3)
-#define as_uint4(x) __builtin_astype((x),  uint4)
-#define as_uint8(x) __builtin_astype((x),  uint8)
-#define as_uint16(x) __builtin_astype((x), uint16)
-
-#define as_long(x) __builtin_astype((x),   long)
-#define as_long2(x) __builtin_astype((x),  long2)
-#define as_long3(x) __builtin_astype((x),  long3)
-#define as_long4(x) __builtin_astype((x),  long4)
-#define as_long8(x) __builtin_astype((x),  long8)
-#define as_long16(x) __builtin_astype((x), long16)
-
-#define as_ulong(x) __builtin_astype((x),   ulong)
-#define as_ulong2(x) __builtin_astype((x),  ulong2)
-#define as_ulong3(x) __builtin_astype((x),  ulong3)
-#define as_ulong4(x) __builtin_astype((x),  ulong4)
-#define as_ulong8(x) __builtin_astype((x),  ulong8)
-#define as_ulong16(x) __builtin_astype((x), ulong16)
-
-#define as_float(x) __builtin_astype((x),   float)
-#define as_float2(x) __builtin_astype((x),  float2)
-#define as_float3(x) __builtin_astype((x),  float3)
-#define as_float4(x) __builtin_astype((x),  float4)
-#define as_float8(x) __builtin_astype((x),  float8)
-#define as_float16(x) __builtin_astype((x), float16)
-
-#ifdef cl_khr_fp64
-#define as_double(x) __builtin_astype((x),   double)
-#define as_double2(x) __builtin_astype((x),  double2)
-#define as_double3(x) __builtin_astype((x),  double3)
-#define as_double4(x) __builtin_astype((x),  double4)
-#define as_double8(x) __builtin_astype((x),  double8)
-#define as_double16(x) __builtin_astype((x), double16)
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-#define as_half(x) __builtin_astype((x),   half)
-#define as_half2(x) __builtin_astype((x),  half2)
-#define as_half3(x) __builtin_astype((x),  half3)
-#define as_half4(x) __builtin_astype((x),  half4)
-#define as_half8(x) __builtin_astype((x),  half8)
-#define as_half16(x) __builtin_astype((x), half16)
-#endif //cl_khr_fp16
-
-// OpenCL v1.1 s6.9, v1.2/2.0 s6.10 - Function qualifiers
-
-#define __kernel_exec(X, typen) __kernel \
-	__attribute__((work_group_size_hint(X, 1, 1))) \
-	__attribute__((vec_type_hint(typen)))
-
-#define kernel_exec(X, typen) __kernel \
-	__attribute__((work_group_size_hint(X, 1, 1))) \
-	__attribute__((vec_type_hint(typen)))
-
-// OpenCL v1.1 s6.11.1, v1.2 s6.12.1, v2.0 s6.13.1 - Work-item Functions
-
-/**
- * Returns the number of dimensions in use. This is the
- * value given to the work_dim argument specified in
- * clEnqueueNDRangeKernel.
- * For clEnqueueTask, this returns 1.
- */
-uint __ovld __cnfn get_work_dim(void);
-
-/**
- * Returns the number of global work-items specified for
- * dimension identified by dimindx. This value is given by
- * the global_work_size argument to
- * clEnqueueNDRangeKernel. Valid values of dimindx
- * are 0 to get_work_dim() - 1. For other values of
- * dimindx, get_global_size() returns 1.
- * For clEnqueueTask, this always returns 1.
- */
-size_t __ovld __cnfn get_global_size(uint dimindx);
-
-/**
- * Returns the unique global work-item ID value for
- * dimension identified by dimindx. The global work-item
- * ID specifies the work-item ID based on the number of
- * global work-items specified to execute the kernel. Valid
- * values of dimindx are 0 to get_work_dim() - 1. For
- * other values of dimindx, get_global_id() returns 0.
- * For clEnqueueTask, this returns 0.
- */
-size_t __ovld __cnfn get_global_id(uint dimindx);
-
-/**
- * Returns the number of local work-items specified in
- * dimension identified by dimindx. This value is given by
- * the local_work_size argument to
- * clEnqueueNDRangeKernel if local_work_size is not
- * NULL; otherwise the OpenCL implementation chooses
- * an appropriate local_work_size value which is returned
- * by this function. Valid values of dimindx are 0 to
- * get_work_dim() - 1. For other values of dimindx,
- * get_local_size() returns 1.
- * For clEnqueueTask, this always returns 1.
- */
-size_t __ovld __cnfn get_local_size(uint dimindx);
-
-/**
- * Returns the unique local work-item ID i.e. a work-item
- * within a specific work-group for dimension identified by
- * dimindx. Valid values of dimindx are 0 to
- * get_work_dim() - 1. For other values of dimindx,
- * get_local_id() returns 0.
- * For clEnqueueTask, this returns 0.
- */
-size_t __ovld __cnfn get_local_id(uint dimindx);
-
-/**
- * Returns the number of work-groups that will execute a
- * kernel for dimension identified by dimindx.
- * Valid values of dimindx are 0 to get_work_dim() - 1.
- * For other values of dimindx, get_num_groups () returns
- * 1.
- * For clEnqueueTask, this always returns 1.
- */
-size_t __ovld __cnfn get_num_groups(uint dimindx);
-
-/**
- * get_group_id returns the work-group ID which is a
- * number from 0 .. get_num_groups(dimindx) - 1.
- * Valid values of dimindx are 0 to get_work_dim() - 1.
- * For other values, get_group_id() returns 0.
- * For clEnqueueTask, this returns 0.
- */
-size_t __ovld __cnfn get_group_id(uint dimindx);
-
-/**
- * get_global_offset returns the offset values specified in
- * global_work_offset argument to
- * clEnqueueNDRangeKernel.
- * Valid values of dimindx are 0 to get_work_dim() - 1.
- * For other values, get_global_offset() returns 0.
- * For clEnqueueTask, this returns 0.
- */
-size_t __ovld __cnfn get_global_offset(uint dimindx);
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-size_t __ovld get_enqueued_local_size(uint dimindx);
-size_t __ovld get_global_linear_id(void);
-size_t __ovld get_local_linear_id(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.2, v1.2 s6.12.2, v2.0 s6.13.2 - Math functions
-
-/**
- * Arc cosine function.
- */
-float __ovld __cnfn acos(float);
-float2 __ovld __cnfn acos(float2);
-float3 __ovld __cnfn acos(float3);
-float4 __ovld __cnfn acos(float4);
-float8 __ovld __cnfn acos(float8);
-float16 __ovld __cnfn acos(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn acos(double);
-double2 __ovld __cnfn acos(double2);
-double3 __ovld __cnfn acos(double3);
-double4 __ovld __cnfn acos(double4);
-double8 __ovld __cnfn acos(double8);
-double16 __ovld __cnfn acos(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn acos(half);
-half2 __ovld __cnfn acos(half2);
-half3 __ovld __cnfn acos(half3);
-half4 __ovld __cnfn acos(half4);
-half8 __ovld __cnfn acos(half8);
-half16 __ovld __cnfn acos(half16);
-#endif //cl_khr_fp16
-
-/**
- * Inverse hyperbolic cosine.
- */
-float __ovld __cnfn acosh(float);
-float2 __ovld __cnfn acosh(float2);
-float3 __ovld __cnfn acosh(float3);
-float4 __ovld __cnfn acosh(float4);
-float8 __ovld __cnfn acosh(float8);
-float16 __ovld __cnfn acosh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn acosh(double);
-double2 __ovld __cnfn acosh(double2);
-double3 __ovld __cnfn acosh(double3);
-double4 __ovld __cnfn acosh(double4);
-double8 __ovld __cnfn acosh(double8);
-double16 __ovld __cnfn acosh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn acosh(half);
-half2 __ovld __cnfn acosh(half2);
-half3 __ovld __cnfn acosh(half3);
-half4 __ovld __cnfn acosh(half4);
-half8 __ovld __cnfn acosh(half8);
-half16 __ovld __cnfn acosh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute acos (x) / PI.
- */
-float __ovld __cnfn acospi(float x);
-float2 __ovld __cnfn acospi(float2 x);
-float3 __ovld __cnfn acospi(float3 x);
-float4 __ovld __cnfn acospi(float4 x);
-float8 __ovld __cnfn acospi(float8 x);
-float16 __ovld __cnfn acospi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn acospi(double x);
-double2 __ovld __cnfn acospi(double2 x);
-double3 __ovld __cnfn acospi(double3 x);
-double4 __ovld __cnfn acospi(double4 x);
-double8 __ovld __cnfn acospi(double8 x);
-double16 __ovld __cnfn acospi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn acospi(half x);
-half2 __ovld __cnfn acospi(half2 x);
-half3 __ovld __cnfn acospi(half3 x);
-half4 __ovld __cnfn acospi(half4 x);
-half8 __ovld __cnfn acospi(half8 x);
-half16 __ovld __cnfn acospi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Arc sine function.
- */
-float __ovld __cnfn asin(float);
-float2 __ovld __cnfn asin(float2);
-float3 __ovld __cnfn asin(float3);
-float4 __ovld __cnfn asin(float4);
-float8 __ovld __cnfn asin(float8);
-float16 __ovld __cnfn asin(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn asin(double);
-double2 __ovld __cnfn asin(double2);
-double3 __ovld __cnfn asin(double3);
-double4 __ovld __cnfn asin(double4);
-double8 __ovld __cnfn asin(double8);
-double16 __ovld __cnfn asin(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn asin(half);
-half2 __ovld __cnfn asin(half2);
-half3 __ovld __cnfn asin(half3);
-half4 __ovld __cnfn asin(half4);
-half8 __ovld __cnfn asin(half8);
-half16 __ovld __cnfn asin(half16);
-#endif //cl_khr_fp16
-
-/**
- * Inverse hyperbolic sine.
- */
-float __ovld __cnfn asinh(float);
-float2 __ovld __cnfn asinh(float2);
-float3 __ovld __cnfn asinh(float3);
-float4 __ovld __cnfn asinh(float4);
-float8 __ovld __cnfn asinh(float8);
-float16 __ovld __cnfn asinh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn asinh(double);
-double2 __ovld __cnfn asinh(double2);
-double3 __ovld __cnfn asinh(double3);
-double4 __ovld __cnfn asinh(double4);
-double8 __ovld __cnfn asinh(double8);
-double16 __ovld __cnfn asinh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn asinh(half);
-half2 __ovld __cnfn asinh(half2);
-half3 __ovld __cnfn asinh(half3);
-half4 __ovld __cnfn asinh(half4);
-half8 __ovld __cnfn asinh(half8);
-half16 __ovld __cnfn asinh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute asin (x) / PI.
- */
-float __ovld __cnfn asinpi(float x);
-float2 __ovld __cnfn asinpi(float2 x);
-float3 __ovld __cnfn asinpi(float3 x);
-float4 __ovld __cnfn asinpi(float4 x);
-float8 __ovld __cnfn asinpi(float8 x);
-float16 __ovld __cnfn asinpi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn asinpi(double x);
-double2 __ovld __cnfn asinpi(double2 x);
-double3 __ovld __cnfn asinpi(double3 x);
-double4 __ovld __cnfn asinpi(double4 x);
-double8 __ovld __cnfn asinpi(double8 x);
-double16 __ovld __cnfn asinpi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn asinpi(half x);
-half2 __ovld __cnfn asinpi(half2 x);
-half3 __ovld __cnfn asinpi(half3 x);
-half4 __ovld __cnfn asinpi(half4 x);
-half8 __ovld __cnfn asinpi(half8 x);
-half16 __ovld __cnfn asinpi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Arc tangent function.
- */
-float __ovld __cnfn atan(float y_over_x);
-float2 __ovld __cnfn atan(float2 y_over_x);
-float3 __ovld __cnfn atan(float3 y_over_x);
-float4 __ovld __cnfn atan(float4 y_over_x);
-float8 __ovld __cnfn atan(float8 y_over_x);
-float16 __ovld __cnfn atan(float16 y_over_x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn atan(double y_over_x);
-double2 __ovld __cnfn atan(double2 y_over_x);
-double3 __ovld __cnfn atan(double3 y_over_x);
-double4 __ovld __cnfn atan(double4 y_over_x);
-double8 __ovld __cnfn atan(double8 y_over_x);
-double16 __ovld __cnfn atan(double16 y_over_x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn atan(half y_over_x);
-half2 __ovld __cnfn atan(half2 y_over_x);
-half3 __ovld __cnfn atan(half3 y_over_x);
-half4 __ovld __cnfn atan(half4 y_over_x);
-half8 __ovld __cnfn atan(half8 y_over_x);
-half16 __ovld __cnfn atan(half16 y_over_x);
-#endif //cl_khr_fp16
-
-/**
- * Arc tangent of y / x.
- */
-float __ovld __cnfn atan2(float y, float x);
-float2 __ovld __cnfn atan2(float2 y, float2 x);
-float3 __ovld __cnfn atan2(float3 y, float3 x);
-float4 __ovld __cnfn atan2(float4 y, float4 x);
-float8 __ovld __cnfn atan2(float8 y, float8 x);
-float16 __ovld __cnfn atan2(float16 y, float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn atan2(double y, double x);
-double2 __ovld __cnfn atan2(double2 y, double2 x);
-double3 __ovld __cnfn atan2(double3 y, double3 x);
-double4 __ovld __cnfn atan2(double4 y, double4 x);
-double8 __ovld __cnfn atan2(double8 y, double8 x);
-double16 __ovld __cnfn atan2(double16 y, double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn atan2(half y, half x);
-half2 __ovld __cnfn atan2(half2 y, half2 x);
-half3 __ovld __cnfn atan2(half3 y, half3 x);
-half4 __ovld __cnfn atan2(half4 y, half4 x);
-half8 __ovld __cnfn atan2(half8 y, half8 x);
-half16 __ovld __cnfn atan2(half16 y, half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Hyperbolic arc tangent.
- */
-float __ovld __cnfn atanh(float);
-float2 __ovld __cnfn atanh(float2);
-float3 __ovld __cnfn atanh(float3);
-float4 __ovld __cnfn atanh(float4);
-float8 __ovld __cnfn atanh(float8);
-float16 __ovld __cnfn atanh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn atanh(double);
-double2 __ovld __cnfn atanh(double2);
-double3 __ovld __cnfn atanh(double3);
-double4 __ovld __cnfn atanh(double4);
-double8 __ovld __cnfn atanh(double8);
-double16 __ovld __cnfn atanh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn atanh(half);
-half2 __ovld __cnfn atanh(half2);
-half3 __ovld __cnfn atanh(half3);
-half4 __ovld __cnfn atanh(half4);
-half8 __ovld __cnfn atanh(half8);
-half16 __ovld __cnfn atanh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute atan (x) / PI.
- */
-float __ovld __cnfn atanpi(float x);
-float2 __ovld __cnfn atanpi(float2 x);
-float3 __ovld __cnfn atanpi(float3 x);
-float4 __ovld __cnfn atanpi(float4 x);
-float8 __ovld __cnfn atanpi(float8 x);
-float16 __ovld __cnfn atanpi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn atanpi(double x);
-double2 __ovld __cnfn atanpi(double2 x);
-double3 __ovld __cnfn atanpi(double3 x);
-double4 __ovld __cnfn atanpi(double4 x);
-double8 __ovld __cnfn atanpi(double8 x);
-double16 __ovld __cnfn atanpi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn atanpi(half x);
-half2 __ovld __cnfn atanpi(half2 x);
-half3 __ovld __cnfn atanpi(half3 x);
-half4 __ovld __cnfn atanpi(half4 x);
-half8 __ovld __cnfn atanpi(half8 x);
-half16 __ovld __cnfn atanpi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute atan2 (y, x) / PI.
- */
-float __ovld __cnfn atan2pi(float y, float x);
-float2 __ovld __cnfn atan2pi(float2 y, float2 x);
-float3 __ovld __cnfn atan2pi(float3 y, float3 x);
-float4 __ovld __cnfn atan2pi(float4 y, float4 x);
-float8 __ovld __cnfn atan2pi(float8 y, float8 x);
-float16 __ovld __cnfn atan2pi(float16 y, float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn atan2pi(double y, double x);
-double2 __ovld __cnfn atan2pi(double2 y, double2 x);
-double3 __ovld __cnfn atan2pi(double3 y, double3 x);
-double4 __ovld __cnfn atan2pi(double4 y, double4 x);
-double8 __ovld __cnfn atan2pi(double8 y, double8 x);
-double16 __ovld __cnfn atan2pi(double16 y, double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn atan2pi(half y, half x);
-half2 __ovld __cnfn atan2pi(half2 y, half2 x);
-half3 __ovld __cnfn atan2pi(half3 y, half3 x);
-half4 __ovld __cnfn atan2pi(half4 y, half4 x);
-half8 __ovld __cnfn atan2pi(half8 y, half8 x);
-half16 __ovld __cnfn atan2pi(half16 y, half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute cube-root.
- */
-float __ovld __cnfn cbrt(float);
-float2 __ovld __cnfn cbrt(float2);
-float3 __ovld __cnfn cbrt(float3);
-float4 __ovld __cnfn cbrt(float4);
-float8 __ovld __cnfn cbrt(float8);
-float16 __ovld __cnfn cbrt(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn cbrt(double);
-double2 __ovld __cnfn cbrt(double2);
-double3 __ovld __cnfn cbrt(double3);
-double4 __ovld __cnfn cbrt(double4);
-double8 __ovld __cnfn cbrt(double8);
-double16 __ovld __cnfn cbrt(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn cbrt(half);
-half2 __ovld __cnfn cbrt(half2);
-half3 __ovld __cnfn cbrt(half3);
-half4 __ovld __cnfn cbrt(half4);
-half8 __ovld __cnfn cbrt(half8);
-half16 __ovld __cnfn cbrt(half16);
-#endif //cl_khr_fp16
-
-/**
- * Round to integral value using the round to positive
- * infinity rounding mode.
- */
-float __ovld __cnfn ceil(float);
-float2 __ovld __cnfn ceil(float2);
-float3 __ovld __cnfn ceil(float3);
-float4 __ovld __cnfn ceil(float4);
-float8 __ovld __cnfn ceil(float8);
-float16 __ovld __cnfn ceil(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn ceil(double);
-double2 __ovld __cnfn ceil(double2);
-double3 __ovld __cnfn ceil(double3);
-double4 __ovld __cnfn ceil(double4);
-double8 __ovld __cnfn ceil(double8);
-double16 __ovld __cnfn ceil(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn ceil(half);
-half2 __ovld __cnfn ceil(half2);
-half3 __ovld __cnfn ceil(half3);
-half4 __ovld __cnfn ceil(half4);
-half8 __ovld __cnfn ceil(half8);
-half16 __ovld __cnfn ceil(half16);
-#endif //cl_khr_fp16
-
-/**
- * Returns x with its sign changed to match the sign of y.
- */
-float __ovld __cnfn copysign(float x, float y);
-float2 __ovld __cnfn copysign(float2 x, float2 y);
-float3 __ovld __cnfn copysign(float3 x, float3 y);
-float4 __ovld __cnfn copysign(float4 x, float4 y);
-float8 __ovld __cnfn copysign(float8 x, float8 y);
-float16 __ovld __cnfn copysign(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn copysign(double x, double y);
-double2 __ovld __cnfn copysign(double2 x, double2 y);
-double3 __ovld __cnfn copysign(double3 x, double3 y);
-double4 __ovld __cnfn copysign(double4 x, double4 y);
-double8 __ovld __cnfn copysign(double8 x, double8 y);
-double16 __ovld __cnfn copysign(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn copysign(half x, half y);
-half2 __ovld __cnfn copysign(half2 x, half2 y);
-half3 __ovld __cnfn copysign(half3 x, half3 y);
-half4 __ovld __cnfn copysign(half4 x, half4 y);
-half8 __ovld __cnfn copysign(half8 x, half8 y);
-half16 __ovld __cnfn copysign(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Compute cosine.
- */
-float __ovld __cnfn cos(float);
-float2 __ovld __cnfn cos(float2);
-float3 __ovld __cnfn cos(float3);
-float4 __ovld __cnfn cos(float4);
-float8 __ovld __cnfn cos(float8);
-float16 __ovld __cnfn cos(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn cos(double);
-double2 __ovld __cnfn cos(double2);
-double3 __ovld __cnfn cos(double3);
-double4 __ovld __cnfn cos(double4);
-double8 __ovld __cnfn cos(double8);
-double16 __ovld __cnfn cos(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn cos(half);
-half2 __ovld __cnfn cos(half2);
-half3 __ovld __cnfn cos(half3);
-half4 __ovld __cnfn cos(half4);
-half8 __ovld __cnfn cos(half8);
-half16 __ovld __cnfn cos(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute hyperbolic cosine.
- */
-float __ovld __cnfn cosh(float);
-float2 __ovld __cnfn cosh(float2);
-float3 __ovld __cnfn cosh(float3);
-float4 __ovld __cnfn cosh(float4);
-float8 __ovld __cnfn cosh(float8);
-float16 __ovld __cnfn cosh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn cosh(double);
-double2 __ovld __cnfn cosh(double2);
-double3 __ovld __cnfn cosh(double3);
-double4 __ovld __cnfn cosh(double4);
-double8 __ovld __cnfn cosh(double8);
-double16 __ovld __cnfn cosh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn cosh(half);
-half2 __ovld __cnfn cosh(half2);
-half3 __ovld __cnfn cosh(half3);
-half4 __ovld __cnfn cosh(half4);
-half8 __ovld __cnfn cosh(half8);
-half16 __ovld __cnfn cosh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute cos (PI * x).
- */
-float __ovld __cnfn cospi(float x);
-float2 __ovld __cnfn cospi(float2 x);
-float3 __ovld __cnfn cospi(float3 x);
-float4 __ovld __cnfn cospi(float4 x);
-float8 __ovld __cnfn cospi(float8 x);
-float16 __ovld __cnfn cospi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn cospi(double x);
-double2 __ovld __cnfn cospi(double2 x);
-double3 __ovld __cnfn cospi(double3 x);
-double4 __ovld __cnfn cospi(double4 x);
-double8 __ovld __cnfn cospi(double8 x);
-double16 __ovld __cnfn cospi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn cospi(half x);
-half2 __ovld __cnfn cospi(half2 x);
-half3 __ovld __cnfn cospi(half3 x);
-half4 __ovld __cnfn cospi(half4 x);
-half8 __ovld __cnfn cospi(half8 x);
-half16 __ovld __cnfn cospi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Complementary error function.
- */
-float __ovld __cnfn erfc(float);
-float2 __ovld __cnfn erfc(float2);
-float3 __ovld __cnfn erfc(float3);
-float4 __ovld __cnfn erfc(float4);
-float8 __ovld __cnfn erfc(float8);
-float16 __ovld __cnfn erfc(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn erfc(double);
-double2 __ovld __cnfn erfc(double2);
-double3 __ovld __cnfn erfc(double3);
-double4 __ovld __cnfn erfc(double4);
-double8 __ovld __cnfn erfc(double8);
-double16 __ovld __cnfn erfc(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn erfc(half);
-half2 __ovld __cnfn erfc(half2);
-half3 __ovld __cnfn erfc(half3);
-half4 __ovld __cnfn erfc(half4);
-half8 __ovld __cnfn erfc(half8);
-half16 __ovld __cnfn erfc(half16);
-#endif //cl_khr_fp16
-
-/**
- * Error function encountered in integrating the
- * normal distribution.
- */
-float __ovld __cnfn erf(float);
-float2 __ovld __cnfn erf(float2);
-float3 __ovld __cnfn erf(float3);
-float4 __ovld __cnfn erf(float4);
-float8 __ovld __cnfn erf(float8);
-float16 __ovld __cnfn erf(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn erf(double);
-double2 __ovld __cnfn erf(double2);
-double3 __ovld __cnfn erf(double3);
-double4 __ovld __cnfn erf(double4);
-double8 __ovld __cnfn erf(double8);
-double16 __ovld __cnfn erf(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn erf(half);
-half2 __ovld __cnfn erf(half2);
-half3 __ovld __cnfn erf(half3);
-half4 __ovld __cnfn erf(half4);
-half8 __ovld __cnfn erf(half8);
-half16 __ovld __cnfn erf(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute the base e exponential function of x.
- */
-float __ovld __cnfn exp(float x);
-float2 __ovld __cnfn exp(float2 x);
-float3 __ovld __cnfn exp(float3 x);
-float4 __ovld __cnfn exp(float4 x);
-float8 __ovld __cnfn exp(float8 x);
-float16 __ovld __cnfn exp(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn exp(double x);
-double2 __ovld __cnfn exp(double2 x);
-double3 __ovld __cnfn exp(double3 x);
-double4 __ovld __cnfn exp(double4 x);
-double8 __ovld __cnfn exp(double8 x);
-double16 __ovld __cnfn exp(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn exp(half x);
-half2 __ovld __cnfn exp(half2 x);
-half3 __ovld __cnfn exp(half3 x);
-half4 __ovld __cnfn exp(half4 x);
-half8 __ovld __cnfn exp(half8 x);
-half16 __ovld __cnfn exp(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Exponential base 2 function.
- */
-float __ovld __cnfn exp2(float);
-float2 __ovld __cnfn exp2(float2);
-float3 __ovld __cnfn exp2(float3);
-float4 __ovld __cnfn exp2(float4);
-float8 __ovld __cnfn exp2(float8);
-float16 __ovld __cnfn exp2(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn exp2(double);
-double2 __ovld __cnfn exp2(double2);
-double3 __ovld __cnfn exp2(double3);
-double4 __ovld __cnfn exp2(double4);
-double8 __ovld __cnfn exp2(double8);
-double16 __ovld __cnfn exp2(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn exp2(half);
-half2 __ovld __cnfn exp2(half2);
-half3 __ovld __cnfn exp2(half3);
-half4 __ovld __cnfn exp2(half4);
-half8 __ovld __cnfn exp2(half8);
-half16 __ovld __cnfn exp2(half16);
-#endif //cl_khr_fp16
-
-/**
- * Exponential base 10 function.
- */
-float __ovld __cnfn exp10(float);
-float2 __ovld __cnfn exp10(float2);
-float3 __ovld __cnfn exp10(float3);
-float4 __ovld __cnfn exp10(float4);
-float8 __ovld __cnfn exp10(float8);
-float16 __ovld __cnfn exp10(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn exp10(double);
-double2 __ovld __cnfn exp10(double2);
-double3 __ovld __cnfn exp10(double3);
-double4 __ovld __cnfn exp10(double4);
-double8 __ovld __cnfn exp10(double8);
-double16 __ovld __cnfn exp10(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn exp10(half);
-half2 __ovld __cnfn exp10(half2);
-half3 __ovld __cnfn exp10(half3);
-half4 __ovld __cnfn exp10(half4);
-half8 __ovld __cnfn exp10(half8);
-half16 __ovld __cnfn exp10(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute e^x- 1.0.
- */
-float __ovld __cnfn expm1(float x);
-float2 __ovld __cnfn expm1(float2 x);
-float3 __ovld __cnfn expm1(float3 x);
-float4 __ovld __cnfn expm1(float4 x);
-float8 __ovld __cnfn expm1(float8 x);
-float16 __ovld __cnfn expm1(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn expm1(double x);
-double2 __ovld __cnfn expm1(double2 x);
-double3 __ovld __cnfn expm1(double3 x);
-double4 __ovld __cnfn expm1(double4 x);
-double8 __ovld __cnfn expm1(double8 x);
-double16 __ovld __cnfn expm1(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn expm1(half x);
-half2 __ovld __cnfn expm1(half2 x);
-half3 __ovld __cnfn expm1(half3 x);
-half4 __ovld __cnfn expm1(half4 x);
-half8 __ovld __cnfn expm1(half8 x);
-half16 __ovld __cnfn expm1(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute absolute value of a floating-point number.
- */
-float __ovld __cnfn fabs(float);
-float2 __ovld __cnfn fabs(float2);
-float3 __ovld __cnfn fabs(float3);
-float4 __ovld __cnfn fabs(float4);
-float8 __ovld __cnfn fabs(float8);
-float16 __ovld __cnfn fabs(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fabs(double);
-double2 __ovld __cnfn fabs(double2);
-double3 __ovld __cnfn fabs(double3);
-double4 __ovld __cnfn fabs(double4);
-double8 __ovld __cnfn fabs(double8);
-double16 __ovld __cnfn fabs(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fabs(half);
-half2 __ovld __cnfn fabs(half2);
-half3 __ovld __cnfn fabs(half3);
-half4 __ovld __cnfn fabs(half4);
-half8 __ovld __cnfn fabs(half8);
-half16 __ovld __cnfn fabs(half16);
-#endif //cl_khr_fp16
-
-/**
- * x - y if x > y, +0 if x is less than or equal to y.
- */
-float __ovld __cnfn fdim(float x, float y);
-float2 __ovld __cnfn fdim(float2 x, float2 y);
-float3 __ovld __cnfn fdim(float3 x, float3 y);
-float4 __ovld __cnfn fdim(float4 x, float4 y);
-float8 __ovld __cnfn fdim(float8 x, float8 y);
-float16 __ovld __cnfn fdim(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fdim(double x, double y);
-double2 __ovld __cnfn fdim(double2 x, double2 y);
-double3 __ovld __cnfn fdim(double3 x, double3 y);
-double4 __ovld __cnfn fdim(double4 x, double4 y);
-double8 __ovld __cnfn fdim(double8 x, double8 y);
-double16 __ovld __cnfn fdim(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fdim(half x, half y);
-half2 __ovld __cnfn fdim(half2 x, half2 y);
-half3 __ovld __cnfn fdim(half3 x, half3 y);
-half4 __ovld __cnfn fdim(half4 x, half4 y);
-half8 __ovld __cnfn fdim(half8 x, half8 y);
-half16 __ovld __cnfn fdim(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Round to integral value using the round to -ve
- * infinity rounding mode.
- */
-float __ovld __cnfn floor(float);
-float2 __ovld __cnfn floor(float2);
-float3 __ovld __cnfn floor(float3);
-float4 __ovld __cnfn floor(float4);
-float8 __ovld __cnfn floor(float8);
-float16 __ovld __cnfn floor(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn floor(double);
-double2 __ovld __cnfn floor(double2);
-double3 __ovld __cnfn floor(double3);
-double4 __ovld __cnfn floor(double4);
-double8 __ovld __cnfn floor(double8);
-double16 __ovld __cnfn floor(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn floor(half);
-half2 __ovld __cnfn floor(half2);
-half3 __ovld __cnfn floor(half3);
-half4 __ovld __cnfn floor(half4);
-half8 __ovld __cnfn floor(half8);
-half16 __ovld __cnfn floor(half16);
-#endif //cl_khr_fp16
-
-/**
- * Returns the correctly rounded floating-point
- * representation of the sum of c with the infinitely
- * precise product of a and b. Rounding of
- * intermediate products shall not occur. Edge case
- * behavior is per the IEEE 754-2008 standard.
- */
-float __ovld __cnfn fma(float a, float b, float c);
-float2 __ovld __cnfn fma(float2 a, float2 b, float2 c);
-float3 __ovld __cnfn fma(float3 a, float3 b, float3 c);
-float4 __ovld __cnfn fma(float4 a, float4 b, float4 c);
-float8 __ovld __cnfn fma(float8 a, float8 b, float8 c);
-float16 __ovld __cnfn fma(float16 a, float16 b, float16 c);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fma(double a, double b, double c);
-double2 __ovld __cnfn fma(double2 a, double2 b, double2 c);
-double3 __ovld __cnfn fma(double3 a, double3 b, double3 c);
-double4 __ovld __cnfn fma(double4 a, double4 b, double4 c);
-double8 __ovld __cnfn fma(double8 a, double8 b, double8 c);
-double16 __ovld __cnfn fma(double16 a, double16 b, double16 c);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fma(half a, half b, half c);
-half2 __ovld __cnfn fma(half2 a, half2 b, half2 c);
-half3 __ovld __cnfn fma(half3 a, half3 b, half3 c);
-half4 __ovld __cnfn fma(half4 a, half4 b, half4 c);
-half8 __ovld __cnfn fma(half8 a, half8 b, half8 c);
-half16 __ovld __cnfn fma(half16 a, half16 b, half16 c);
-#endif //cl_khr_fp16
-
-/**
- * Returns y if x < y, otherwise it returns x. If one
- * argument is a NaN, fmax() returns the other
- * argument. If both arguments are NaNs, fmax()
- * returns a NaN.
- */
-float __ovld __cnfn fmax(float x, float y);
-float2 __ovld __cnfn fmax(float2 x, float2 y);
-float3 __ovld __cnfn fmax(float3 x, float3 y);
-float4 __ovld __cnfn fmax(float4 x, float4 y);
-float8 __ovld __cnfn fmax(float8 x, float8 y);
-float16 __ovld __cnfn fmax(float16 x, float16 y);
-float2 __ovld __cnfn fmax(float2 x, float y);
-float3 __ovld __cnfn fmax(float3 x, float y);
-float4 __ovld __cnfn fmax(float4 x, float y);
-float8 __ovld __cnfn fmax(float8 x, float y);
-float16 __ovld __cnfn fmax(float16 x, float y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fmax(double x, double y);
-double2 __ovld __cnfn fmax(double2 x, double2 y);
-double3 __ovld __cnfn fmax(double3 x, double3 y);
-double4 __ovld __cnfn fmax(double4 x, double4 y);
-double8 __ovld __cnfn fmax(double8 x, double8 y);
-double16 __ovld __cnfn fmax(double16 x, double16 y);
-double2 __ovld __cnfn fmax(double2 x, double y);
-double3 __ovld __cnfn fmax(double3 x, double y);
-double4 __ovld __cnfn fmax(double4 x, double y);
-double8 __ovld __cnfn fmax(double8 x, double y);
-double16 __ovld __cnfn fmax(double16 x, double y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fmax(half x, half y);
-half2 __ovld __cnfn fmax(half2 x, half2 y);
-half3 __ovld __cnfn fmax(half3 x, half3 y);
-half4 __ovld __cnfn fmax(half4 x, half4 y);
-half8 __ovld __cnfn fmax(half8 x, half8 y);
-half16 __ovld __cnfn fmax(half16 x, half16 y);
-half2 __ovld __cnfn fmax(half2 x, half y);
-half3 __ovld __cnfn fmax(half3 x, half y);
-half4 __ovld __cnfn fmax(half4 x, half y);
-half8 __ovld __cnfn fmax(half8 x, half y);
-half16 __ovld __cnfn fmax(half16 x, half y);
-#endif //cl_khr_fp16
-
-/**
- * Returns y if y < x, otherwise it returns x. If one
- * argument is a NaN, fmin() returns the other
- * argument. If both arguments are NaNs, fmin()
- * returns a NaN.
- */
-float __ovld __cnfn fmin(float x, float y);
-float2 __ovld __cnfn fmin(float2 x, float2 y);
-float3 __ovld __cnfn fmin(float3 x, float3 y);
-float4 __ovld __cnfn fmin(float4 x, float4 y);
-float8 __ovld __cnfn fmin(float8 x, float8 y);
-float16 __ovld __cnfn fmin(float16 x, float16 y);
-float2 __ovld __cnfn fmin(float2 x, float y);
-float3 __ovld __cnfn fmin(float3 x, float y);
-float4 __ovld __cnfn fmin(float4 x, float y);
-float8 __ovld __cnfn fmin(float8 x, float y);
-float16 __ovld __cnfn fmin(float16 x, float y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fmin(double x, double y);
-double2 __ovld __cnfn fmin(double2 x, double2 y);
-double3 __ovld __cnfn fmin(double3 x, double3 y);
-double4 __ovld __cnfn fmin(double4 x, double4 y);
-double8 __ovld __cnfn fmin(double8 x, double8 y);
-double16 __ovld __cnfn fmin(double16 x, double16 y);
-double2 __ovld __cnfn fmin(double2 x, double y);
-double3 __ovld __cnfn fmin(double3 x, double y);
-double4 __ovld __cnfn fmin(double4 x, double y);
-double8 __ovld __cnfn fmin(double8 x, double y);
-double16 __ovld __cnfn fmin(double16 x, double y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fmin(half x, half y);
-half2 __ovld __cnfn fmin(half2 x, half2 y);
-half3 __ovld __cnfn fmin(half3 x, half3 y);
-half4 __ovld __cnfn fmin(half4 x, half4 y);
-half8 __ovld __cnfn fmin(half8 x, half8 y);
-half16 __ovld __cnfn fmin(half16 x, half16 y);
-half2 __ovld __cnfn fmin(half2 x, half y);
-half3 __ovld __cnfn fmin(half3 x, half y);
-half4 __ovld __cnfn fmin(half4 x, half y);
-half8 __ovld __cnfn fmin(half8 x, half y);
-half16 __ovld __cnfn fmin(half16 x, half y);
-#endif //cl_khr_fp16
-
-/**
- * Modulus. Returns x - y * trunc (x/y).
- */
-float __ovld __cnfn fmod(float x, float y);
-float2 __ovld __cnfn fmod(float2 x, float2 y);
-float3 __ovld __cnfn fmod(float3 x, float3 y);
-float4 __ovld __cnfn fmod(float4 x, float4 y);
-float8 __ovld __cnfn fmod(float8 x, float8 y);
-float16 __ovld __cnfn fmod(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn fmod(double x, double y);
-double2 __ovld __cnfn fmod(double2 x, double2 y);
-double3 __ovld __cnfn fmod(double3 x, double3 y);
-double4 __ovld __cnfn fmod(double4 x, double4 y);
-double8 __ovld __cnfn fmod(double8 x, double8 y);
-double16 __ovld __cnfn fmod(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn fmod(half x, half y);
-half2 __ovld __cnfn fmod(half2 x, half2 y);
-half3 __ovld __cnfn fmod(half3 x, half3 y);
-half4 __ovld __cnfn fmod(half4 x, half4 y);
-half8 __ovld __cnfn fmod(half8 x, half8 y);
-half16 __ovld __cnfn fmod(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns fmin(x - floor (x), 0x1.fffffep-1f ).
- * floor(x) is returned in iptr.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld fract(float x, float *iptr);
-float2 __ovld fract(float2 x, float2 *iptr);
-float3 __ovld fract(float3 x, float3 *iptr);
-float4 __ovld fract(float4 x, float4 *iptr);
-float8 __ovld fract(float8 x, float8 *iptr);
-float16 __ovld fract(float16 x, float16 *iptr);
-#ifdef cl_khr_fp64
-double __ovld fract(double x, double *iptr);
-double2 __ovld fract(double2 x, double2 *iptr);
-double3 __ovld fract(double3 x, double3 *iptr);
-double4 __ovld fract(double4 x, double4 *iptr);
-double8 __ovld fract(double8 x, double8 *iptr);
-double16 __ovld fract(double16 x, double16 *iptr);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld fract(half x, half *iptr);
-half2 __ovld fract(half2 x, half2 *iptr);
-half3 __ovld fract(half3 x, half3 *iptr);
-half4 __ovld fract(half4 x, half4 *iptr);
-half8 __ovld fract(half8 x, half8 *iptr);
-half16 __ovld fract(half16 x, half16 *iptr);
-#endif //cl_khr_fp16
-#else
-float __ovld fract(float x, __global float *iptr);
-float2 __ovld fract(float2 x, __global float2 *iptr);
-float3 __ovld fract(float3 x, __global float3 *iptr);
-float4 __ovld fract(float4 x, __global float4 *iptr);
-float8 __ovld fract(float8 x, __global float8 *iptr);
-float16 __ovld fract(float16 x, __global float16 *iptr);
-float __ovld fract(float x, __local float *iptr);
-float2 __ovld fract(float2 x, __local float2 *iptr);
-float3 __ovld fract(float3 x, __local float3 *iptr);
-float4 __ovld fract(float4 x, __local float4 *iptr);
-float8 __ovld fract(float8 x, __local float8 *iptr);
-float16 __ovld fract(float16 x, __local float16 *iptr);
-float __ovld fract(float x, __private float *iptr);
-float2 __ovld fract(float2 x, __private float2 *iptr);
-float3 __ovld fract(float3 x, __private float3 *iptr);
-float4 __ovld fract(float4 x, __private float4 *iptr);
-float8 __ovld fract(float8 x, __private float8 *iptr);
-float16 __ovld fract(float16 x, __private float16 *iptr);
-#ifdef cl_khr_fp64
-double __ovld fract(double x, __global double *iptr);
-double2 __ovld fract(double2 x, __global double2 *iptr);
-double3 __ovld fract(double3 x, __global double3 *iptr);
-double4 __ovld fract(double4 x, __global double4 *iptr);
-double8 __ovld fract(double8 x, __global double8 *iptr);
-double16 __ovld fract(double16 x, __global double16 *iptr);
-double __ovld fract(double x, __local double *iptr);
-double2 __ovld fract(double2 x, __local double2 *iptr);
-double3 __ovld fract(double3 x, __local double3 *iptr);
-double4 __ovld fract(double4 x, __local double4 *iptr);
-double8 __ovld fract(double8 x, __local double8 *iptr);
-double16 __ovld fract(double16 x, __local double16 *iptr);
-double __ovld fract(double x, __private double *iptr);
-double2 __ovld fract(double2 x, __private double2 *iptr);
-double3 __ovld fract(double3 x, __private double3 *iptr);
-double4 __ovld fract(double4 x, __private double4 *iptr);
-double8 __ovld fract(double8 x, __private double8 *iptr);
-double16 __ovld fract(double16 x, __private double16 *iptr);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld fract(half x, __global half *iptr);
-half2 __ovld fract(half2 x, __global half2 *iptr);
-half3 __ovld fract(half3 x, __global half3 *iptr);
-half4 __ovld fract(half4 x, __global half4 *iptr);
-half8 __ovld fract(half8 x, __global half8 *iptr);
-half16 __ovld fract(half16 x, __global half16 *iptr);
-half __ovld fract(half x, __local half *iptr);
-half2 __ovld fract(half2 x, __local half2 *iptr);
-half3 __ovld fract(half3 x, __local half3 *iptr);
-half4 __ovld fract(half4 x, __local half4 *iptr);
-half8 __ovld fract(half8 x, __local half8 *iptr);
-half16 __ovld fract(half16 x, __local half16 *iptr);
-half __ovld fract(half x, __private half *iptr);
-half2 __ovld fract(half2 x, __private half2 *iptr);
-half3 __ovld fract(half3 x, __private half3 *iptr);
-half4 __ovld fract(half4 x, __private half4 *iptr);
-half8 __ovld fract(half8 x, __private half8 *iptr);
-half16 __ovld fract(half16 x, __private half16 *iptr);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Extract mantissa and exponent from x. For each
- * component the mantissa returned is a float with
- * magnitude in the interval [1/2, 1) or 0. Each
- * component of x equals mantissa returned * 2^exp.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld frexp(float x, int *exp);
-float2 __ovld frexp(float2 x, int2 *exp);
-float3 __ovld frexp(float3 x, int3 *exp);
-float4 __ovld frexp(float4 x, int4 *exp);
-float8 __ovld frexp(float8 x, int8 *exp);
-float16 __ovld frexp(float16 x, int16 *exp);
-#ifdef cl_khr_fp64
-double __ovld frexp(double x, int *exp);
-double2 __ovld frexp(double2 x, int2 *exp);
-double3 __ovld frexp(double3 x, int3 *exp);
-double4 __ovld frexp(double4 x, int4 *exp);
-double8 __ovld frexp(double8 x, int8 *exp);
-double16 __ovld frexp(double16 x, int16 *exp);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld frexp(half x, int *exp);
-half2 __ovld frexp(half2 x, int2 *exp);
-half3 __ovld frexp(half3 x, int3 *exp);
-half4 __ovld frexp(half4 x, int4 *exp);
-half8 __ovld frexp(half8 x, int8 *exp);
-half16 __ovld frexp(half16 x, int16 *exp);
-#endif //cl_khr_fp16
-#else
-float __ovld frexp(float x, __global int *exp);
-float2 __ovld frexp(float2 x, __global int2 *exp);
-float3 __ovld frexp(float3 x, __global int3 *exp);
-float4 __ovld frexp(float4 x, __global int4 *exp);
-float8 __ovld frexp(float8 x, __global int8 *exp);
-float16 __ovld frexp(float16 x, __global int16 *exp);
-float __ovld frexp(float x, __local int *exp);
-float2 __ovld frexp(float2 x, __local int2 *exp);
-float3 __ovld frexp(float3 x, __local int3 *exp);
-float4 __ovld frexp(float4 x, __local int4 *exp);
-float8 __ovld frexp(float8 x, __local int8 *exp);
-float16 __ovld frexp(float16 x, __local int16 *exp);
-float __ovld frexp(float x, __private int *exp);
-float2 __ovld frexp(float2 x, __private int2 *exp);
-float3 __ovld frexp(float3 x, __private int3 *exp);
-float4 __ovld frexp(float4 x, __private int4 *exp);
-float8 __ovld frexp(float8 x, __private int8 *exp);
-float16 __ovld frexp(float16 x, __private int16 *exp);
-#ifdef cl_khr_fp64
-double __ovld frexp(double x, __global int *exp);
-double2 __ovld frexp(double2 x, __global int2 *exp);
-double3 __ovld frexp(double3 x, __global int3 *exp);
-double4 __ovld frexp(double4 x, __global int4 *exp);
-double8 __ovld frexp(double8 x, __global int8 *exp);
-double16 __ovld frexp(double16 x, __global int16 *exp);
-double __ovld frexp(double x, __local int *exp);
-double2 __ovld frexp(double2 x, __local int2 *exp);
-double3 __ovld frexp(double3 x, __local int3 *exp);
-double4 __ovld frexp(double4 x, __local int4 *exp);
-double8 __ovld frexp(double8 x, __local int8 *exp);
-double16 __ovld frexp(double16 x, __local int16 *exp);
-double __ovld frexp(double x, __private int *exp);
-double2 __ovld frexp(double2 x, __private int2 *exp);
-double3 __ovld frexp(double3 x, __private int3 *exp);
-double4 __ovld frexp(double4 x, __private int4 *exp);
-double8 __ovld frexp(double8 x, __private int8 *exp);
-double16 __ovld frexp(double16 x, __private int16 *exp);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld frexp(half x, __global int *exp);
-half2 __ovld frexp(half2 x, __global int2 *exp);
-half3 __ovld frexp(half3 x, __global int3 *exp);
-half4 __ovld frexp(half4 x, __global int4 *exp);
-half8 __ovld frexp(half8 x, __global int8 *exp);
-half16 __ovld frexp(half16 x, __global int16 *exp);
-half __ovld frexp(half x, __local int *exp);
-half2 __ovld frexp(half2 x, __local int2 *exp);
-half3 __ovld frexp(half3 x, __local int3 *exp);
-half4 __ovld frexp(half4 x, __local int4 *exp);
-half8 __ovld frexp(half8 x, __local int8 *exp);
-half16 __ovld frexp(half16 x, __local int16 *exp);
-half __ovld frexp(half x, __private int *exp);
-half2 __ovld frexp(half2 x, __private int2 *exp);
-half3 __ovld frexp(half3 x, __private int3 *exp);
-half4 __ovld frexp(half4 x, __private int4 *exp);
-half8 __ovld frexp(half8 x, __private int8 *exp);
-half16 __ovld frexp(half16 x, __private int16 *exp);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Compute the value of the square root of x^2 + y^2
- * without undue overflow or underflow.
- */
-float __ovld __cnfn hypot(float x, float y);
-float2 __ovld __cnfn hypot(float2 x, float2 y);
-float3 __ovld __cnfn hypot(float3 x, float3 y);
-float4 __ovld __cnfn hypot(float4 x, float4 y);
-float8 __ovld __cnfn hypot(float8 x, float8 y);
-float16 __ovld __cnfn hypot(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn hypot(double x, double y);
-double2 __ovld __cnfn hypot(double2 x, double2 y);
-double3 __ovld __cnfn hypot(double3 x, double3 y);
-double4 __ovld __cnfn hypot(double4 x, double4 y);
-double8 __ovld __cnfn hypot(double8 x, double8 y);
-double16 __ovld __cnfn hypot(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn hypot(half x, half y);
-half2 __ovld __cnfn hypot(half2 x, half2 y);
-half3 __ovld __cnfn hypot(half3 x, half3 y);
-half4 __ovld __cnfn hypot(half4 x, half4 y);
-half8 __ovld __cnfn hypot(half8 x, half8 y);
-half16 __ovld __cnfn hypot(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Return the exponent as an integer value.
- */
-int __ovld __cnfn ilogb(float x);
-int2 __ovld __cnfn ilogb(float2 x);
-int3 __ovld __cnfn ilogb(float3 x);
-int4 __ovld __cnfn ilogb(float4 x);
-int8 __ovld __cnfn ilogb(float8 x);
-int16 __ovld __cnfn ilogb(float16 x);
-#ifdef cl_khr_fp64
-int __ovld __cnfn ilogb(double x);
-int2 __ovld __cnfn ilogb(double2 x);
-int3 __ovld __cnfn ilogb(double3 x);
-int4 __ovld __cnfn ilogb(double4 x);
-int8 __ovld __cnfn ilogb(double8 x);
-int16 __ovld __cnfn ilogb(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn ilogb(half x);
-int2 __ovld __cnfn ilogb(half2 x);
-int3 __ovld __cnfn ilogb(half3 x);
-int4 __ovld __cnfn ilogb(half4 x);
-int8 __ovld __cnfn ilogb(half8 x);
-int16 __ovld __cnfn ilogb(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Multiply x by 2 to the power n.
- */
-float __ovld __cnfn ldexp(float x, int n);
-float2 __ovld __cnfn ldexp(float2 x, int2 n);
-float3 __ovld __cnfn ldexp(float3 x, int3 n);
-float4 __ovld __cnfn ldexp(float4 x, int4 n);
-float8 __ovld __cnfn ldexp(float8 x, int8 n);
-float16 __ovld __cnfn ldexp(float16 x, int16 n);
-float2 __ovld __cnfn ldexp(float2 x, int n);
-float3 __ovld __cnfn ldexp(float3 x, int n);
-float4 __ovld __cnfn ldexp(float4 x, int n);
-float8 __ovld __cnfn ldexp(float8 x, int n);
-float16 __ovld __cnfn ldexp(float16 x, int n);
-#ifdef cl_khr_fp64
-double __ovld __cnfn ldexp(double x, int n);
-double2 __ovld __cnfn ldexp(double2 x, int2 n);
-double3 __ovld __cnfn ldexp(double3 x, int3 n);
-double4 __ovld __cnfn ldexp(double4 x, int4 n);
-double8 __ovld __cnfn ldexp(double8 x, int8 n);
-double16 __ovld __cnfn ldexp(double16 x, int16 n);
-double2 __ovld __cnfn ldexp(double2 x, int n);
-double3 __ovld __cnfn ldexp(double3 x, int n);
-double4 __ovld __cnfn ldexp(double4 x, int n);
-double8 __ovld __cnfn ldexp(double8 x, int n);
-double16 __ovld __cnfn ldexp(double16 x, int n);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn ldexp(half x, int n);
-half2 __ovld __cnfn ldexp(half2 x, int2 n);
-half3 __ovld __cnfn ldexp(half3 x, int3 n);
-half4 __ovld __cnfn ldexp(half4 x, int4 n);
-half8 __ovld __cnfn ldexp(half8 x, int8 n);
-half16 __ovld __cnfn ldexp(half16 x, int16 n);
-half2 __ovld __cnfn ldexp(half2 x, int n);
-half3 __ovld __cnfn ldexp(half3 x, int n);
-half4 __ovld __cnfn ldexp(half4 x, int n);
-half8 __ovld __cnfn ldexp(half8 x, int n);
-half16 __ovld __cnfn ldexp(half16 x, int n);
-#endif //cl_khr_fp16
-
-/**
- * Log gamma function. Returns the natural
- * logarithm of the absolute value of the gamma
- * function. The sign of the gamma function is
- * returned in the signp argument of lgamma_r.
- */
-float __ovld __cnfn lgamma(float x);
-float2 __ovld __cnfn lgamma(float2 x);
-float3 __ovld __cnfn lgamma(float3 x);
-float4 __ovld __cnfn lgamma(float4 x);
-float8 __ovld __cnfn lgamma(float8 x);
-float16 __ovld __cnfn lgamma(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn lgamma(double x);
-double2 __ovld __cnfn lgamma(double2 x);
-double3 __ovld __cnfn lgamma(double3 x);
-double4 __ovld __cnfn lgamma(double4 x);
-double8 __ovld __cnfn lgamma(double8 x);
-double16 __ovld __cnfn lgamma(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn lgamma(half x);
-half2 __ovld __cnfn lgamma(half2 x);
-half3 __ovld __cnfn lgamma(half3 x);
-half4 __ovld __cnfn lgamma(half4 x);
-half8 __ovld __cnfn lgamma(half8 x);
-half16 __ovld __cnfn lgamma(half16 x);
-#endif //cl_khr_fp16
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld lgamma_r(float x, int *signp);
-float2 __ovld lgamma_r(float2 x, int2 *signp);
-float3 __ovld lgamma_r(float3 x, int3 *signp);
-float4 __ovld lgamma_r(float4 x, int4 *signp);
-float8 __ovld lgamma_r(float8 x, int8 *signp);
-float16 __ovld lgamma_r(float16 x, int16 *signp);
-#ifdef cl_khr_fp64
-double __ovld lgamma_r(double x, int *signp);
-double2 __ovld lgamma_r(double2 x, int2 *signp);
-double3 __ovld lgamma_r(double3 x, int3 *signp);
-double4 __ovld lgamma_r(double4 x, int4 *signp);
-double8 __ovld lgamma_r(double8 x, int8 *signp);
-double16 __ovld lgamma_r(double16 x, int16 *signp);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld lgamma_r(half x, int *signp);
-half2 __ovld lgamma_r(half2 x, int2 *signp);
-half3 __ovld lgamma_r(half3 x, int3 *signp);
-half4 __ovld lgamma_r(half4 x, int4 *signp);
-half8 __ovld lgamma_r(half8 x, int8 *signp);
-half16 __ovld lgamma_r(half16 x, int16 *signp);
-#endif //cl_khr_fp16
-#else
-float __ovld lgamma_r(float x, __global int *signp);
-float2 __ovld lgamma_r(float2 x, __global int2 *signp);
-float3 __ovld lgamma_r(float3 x, __global int3 *signp);
-float4 __ovld lgamma_r(float4 x, __global int4 *signp);
-float8 __ovld lgamma_r(float8 x, __global int8 *signp);
-float16 __ovld lgamma_r(float16 x, __global int16 *signp);
-float __ovld lgamma_r(float x, __local int *signp);
-float2 __ovld lgamma_r(float2 x, __local int2 *signp);
-float3 __ovld lgamma_r(float3 x, __local int3 *signp);
-float4 __ovld lgamma_r(float4 x, __local int4 *signp);
-float8 __ovld lgamma_r(float8 x, __local int8 *signp);
-float16 __ovld lgamma_r(float16 x, __local int16 *signp);
-float __ovld lgamma_r(float x, __private int *signp);
-float2 __ovld lgamma_r(float2 x, __private int2 *signp);
-float3 __ovld lgamma_r(float3 x, __private int3 *signp);
-float4 __ovld lgamma_r(float4 x, __private int4 *signp);
-float8 __ovld lgamma_r(float8 x, __private int8 *signp);
-float16 __ovld lgamma_r(float16 x, __private int16 *signp);
-#ifdef cl_khr_fp64
-double __ovld lgamma_r(double x, __global int *signp);
-double2 __ovld lgamma_r(double2 x, __global int2 *signp);
-double3 __ovld lgamma_r(double3 x, __global int3 *signp);
-double4 __ovld lgamma_r(double4 x, __global int4 *signp);
-double8 __ovld lgamma_r(double8 x, __global int8 *signp);
-double16 __ovld lgamma_r(double16 x, __global int16 *signp);
-double __ovld lgamma_r(double x, __local int *signp);
-double2 __ovld lgamma_r(double2 x, __local int2 *signp);
-double3 __ovld lgamma_r(double3 x, __local int3 *signp);
-double4 __ovld lgamma_r(double4 x, __local int4 *signp);
-double8 __ovld lgamma_r(double8 x, __local int8 *signp);
-double16 __ovld lgamma_r(double16 x, __local int16 *signp);
-double __ovld lgamma_r(double x, __private int *signp);
-double2 __ovld lgamma_r(double2 x, __private int2 *signp);
-double3 __ovld lgamma_r(double3 x, __private int3 *signp);
-double4 __ovld lgamma_r(double4 x, __private int4 *signp);
-double8 __ovld lgamma_r(double8 x, __private int8 *signp);
-double16 __ovld lgamma_r(double16 x, __private int16 *signp);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld lgamma_r(half x, __global int *signp);
-half2 __ovld lgamma_r(half2 x, __global int2 *signp);
-half3 __ovld lgamma_r(half3 x, __global int3 *signp);
-half4 __ovld lgamma_r(half4 x, __global int4 *signp);
-half8 __ovld lgamma_r(half8 x, __global int8 *signp);
-half16 __ovld lgamma_r(half16 x, __global int16 *signp);
-half __ovld lgamma_r(half x, __local int *signp);
-half2 __ovld lgamma_r(half2 x, __local int2 *signp);
-half3 __ovld lgamma_r(half3 x, __local int3 *signp);
-half4 __ovld lgamma_r(half4 x, __local int4 *signp);
-half8 __ovld lgamma_r(half8 x, __local int8 *signp);
-half16 __ovld lgamma_r(half16 x, __local int16 *signp);
-half __ovld lgamma_r(half x, __private int *signp);
-half2 __ovld lgamma_r(half2 x, __private int2 *signp);
-half3 __ovld lgamma_r(half3 x, __private int3 *signp);
-half4 __ovld lgamma_r(half4 x, __private int4 *signp);
-half8 __ovld lgamma_r(half8 x, __private int8 *signp);
-half16 __ovld lgamma_r(half16 x, __private int16 *signp);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Compute natural logarithm.
- */
-float __ovld __cnfn log(float);
-float2 __ovld __cnfn log(float2);
-float3 __ovld __cnfn log(float3);
-float4 __ovld __cnfn log(float4);
-float8 __ovld __cnfn log(float8);
-float16 __ovld __cnfn log(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn log(double);
-double2 __ovld __cnfn log(double2);
-double3 __ovld __cnfn log(double3);
-double4 __ovld __cnfn log(double4);
-double8 __ovld __cnfn log(double8);
-double16 __ovld __cnfn log(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn log(half);
-half2 __ovld __cnfn log(half2);
-half3 __ovld __cnfn log(half3);
-half4 __ovld __cnfn log(half4);
-half8 __ovld __cnfn log(half8);
-half16 __ovld __cnfn log(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute a base 2 logarithm.
- */
-float __ovld __cnfn log2(float);
-float2 __ovld __cnfn log2(float2);
-float3 __ovld __cnfn log2(float3);
-float4 __ovld __cnfn log2(float4);
-float8 __ovld __cnfn log2(float8);
-float16 __ovld __cnfn log2(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn log2(double);
-double2 __ovld __cnfn log2(double2);
-double3 __ovld __cnfn log2(double3);
-double4 __ovld __cnfn log2(double4);
-double8 __ovld __cnfn log2(double8);
-double16 __ovld __cnfn log2(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn log2(half);
-half2 __ovld __cnfn log2(half2);
-half3 __ovld __cnfn log2(half3);
-half4 __ovld __cnfn log2(half4);
-half8 __ovld __cnfn log2(half8);
-half16 __ovld __cnfn log2(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute a base 10 logarithm.
- */
-float __ovld __cnfn log10(float);
-float2 __ovld __cnfn log10(float2);
-float3 __ovld __cnfn log10(float3);
-float4 __ovld __cnfn log10(float4);
-float8 __ovld __cnfn log10(float8);
-float16 __ovld __cnfn log10(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn log10(double);
-double2 __ovld __cnfn log10(double2);
-double3 __ovld __cnfn log10(double3);
-double4 __ovld __cnfn log10(double4);
-double8 __ovld __cnfn log10(double8);
-double16 __ovld __cnfn log10(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn log10(half);
-half2 __ovld __cnfn log10(half2);
-half3 __ovld __cnfn log10(half3);
-half4 __ovld __cnfn log10(half4);
-half8 __ovld __cnfn log10(half8);
-half16 __ovld __cnfn log10(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute a base e logarithm of (1.0 + x).
- */
-float __ovld __cnfn log1p(float x);
-float2 __ovld __cnfn log1p(float2 x);
-float3 __ovld __cnfn log1p(float3 x);
-float4 __ovld __cnfn log1p(float4 x);
-float8 __ovld __cnfn log1p(float8 x);
-float16 __ovld __cnfn log1p(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn log1p(double x);
-double2 __ovld __cnfn log1p(double2 x);
-double3 __ovld __cnfn log1p(double3 x);
-double4 __ovld __cnfn log1p(double4 x);
-double8 __ovld __cnfn log1p(double8 x);
-double16 __ovld __cnfn log1p(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn log1p(half x);
-half2 __ovld __cnfn log1p(half2 x);
-half3 __ovld __cnfn log1p(half3 x);
-half4 __ovld __cnfn log1p(half4 x);
-half8 __ovld __cnfn log1p(half8 x);
-half16 __ovld __cnfn log1p(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute the exponent of x, which is the integral
- * part of logr | x |.
- */
-float __ovld __cnfn logb(float x);
-float2 __ovld __cnfn logb(float2 x);
-float3 __ovld __cnfn logb(float3 x);
-float4 __ovld __cnfn logb(float4 x);
-float8 __ovld __cnfn logb(float8 x);
-float16 __ovld __cnfn logb(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn logb(double x);
-double2 __ovld __cnfn logb(double2 x);
-double3 __ovld __cnfn logb(double3 x);
-double4 __ovld __cnfn logb(double4 x);
-double8 __ovld __cnfn logb(double8 x);
-double16 __ovld __cnfn logb(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn logb(half x);
-half2 __ovld __cnfn logb(half2 x);
-half3 __ovld __cnfn logb(half3 x);
-half4 __ovld __cnfn logb(half4 x);
-half8 __ovld __cnfn logb(half8 x);
-half16 __ovld __cnfn logb(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * mad approximates a * b + c. Whether or how the
- * product of a * b is rounded and how supernormal or
- * subnormal intermediate products are handled is not
- * defined. mad is intended to be used where speed is
- * preferred over accuracy.
- */
-float __ovld __cnfn mad(float a, float b, float c);
-float2 __ovld __cnfn mad(float2 a, float2 b, float2 c);
-float3 __ovld __cnfn mad(float3 a, float3 b, float3 c);
-float4 __ovld __cnfn mad(float4 a, float4 b, float4 c);
-float8 __ovld __cnfn mad(float8 a, float8 b, float8 c);
-float16 __ovld __cnfn mad(float16 a, float16 b, float16 c);
-#ifdef cl_khr_fp64
-double __ovld __cnfn mad(double a, double b, double c);
-double2 __ovld __cnfn mad(double2 a, double2 b, double2 c);
-double3 __ovld __cnfn mad(double3 a, double3 b, double3 c);
-double4 __ovld __cnfn mad(double4 a, double4 b, double4 c);
-double8 __ovld __cnfn mad(double8 a, double8 b, double8 c);
-double16 __ovld __cnfn mad(double16 a, double16 b, double16 c);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn mad(half a, half b, half c);
-half2 __ovld __cnfn mad(half2 a, half2 b, half2 c);
-half3 __ovld __cnfn mad(half3 a, half3 b, half3 c);
-half4 __ovld __cnfn mad(half4 a, half4 b, half4 c);
-half8 __ovld __cnfn mad(half8 a, half8 b, half8 c);
-half16 __ovld __cnfn mad(half16 a, half16 b, half16 c);
-#endif //cl_khr_fp16
-
-/**
- * Returns x if | x | > | y |, y if | y | > | x |, otherwise
- * fmax(x, y).
- */
-float __ovld __cnfn maxmag(float x, float y);
-float2 __ovld __cnfn maxmag(float2 x, float2 y);
-float3 __ovld __cnfn maxmag(float3 x, float3 y);
-float4 __ovld __cnfn maxmag(float4 x, float4 y);
-float8 __ovld __cnfn maxmag(float8 x, float8 y);
-float16 __ovld __cnfn maxmag(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn maxmag(double x, double y);
-double2 __ovld __cnfn maxmag(double2 x, double2 y);
-double3 __ovld __cnfn maxmag(double3 x, double3 y);
-double4 __ovld __cnfn maxmag(double4 x, double4 y);
-double8 __ovld __cnfn maxmag(double8 x, double8 y);
-double16 __ovld __cnfn maxmag(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn maxmag(half x, half y);
-half2 __ovld __cnfn maxmag(half2 x, half2 y);
-half3 __ovld __cnfn maxmag(half3 x, half3 y);
-half4 __ovld __cnfn maxmag(half4 x, half4 y);
-half8 __ovld __cnfn maxmag(half8 x, half8 y);
-half16 __ovld __cnfn maxmag(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns x if | x | < | y |, y if | y | < | x |, otherwise
- * fmin(x, y).
- */
-float __ovld __cnfn minmag(float x, float y);
-float2 __ovld __cnfn minmag(float2 x, float2 y);
-float3 __ovld __cnfn minmag(float3 x, float3 y);
-float4 __ovld __cnfn minmag(float4 x, float4 y);
-float8 __ovld __cnfn minmag(float8 x, float8 y);
-float16 __ovld __cnfn minmag(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn minmag(double x, double y);
-double2 __ovld __cnfn minmag(double2 x, double2 y);
-double3 __ovld __cnfn minmag(double3 x, double3 y);
-double4 __ovld __cnfn minmag(double4 x, double4 y);
-double8 __ovld __cnfn minmag(double8 x, double8 y);
-double16 __ovld __cnfn minmag(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn minmag(half x, half y);
-half2 __ovld __cnfn minmag(half2 x, half2 y);
-half3 __ovld __cnfn minmag(half3 x, half3 y);
-half4 __ovld __cnfn minmag(half4 x, half4 y);
-half8 __ovld __cnfn minmag(half8 x, half8 y);
-half16 __ovld __cnfn minmag(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Decompose a floating-point number. The modf
- * function breaks the argument x into integral and
- * fractional parts, each of which has the same sign as
- * the argument. It stores the integral part in the object
- * pointed to by iptr.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld modf(float x, float *iptr);
-float2 __ovld modf(float2 x, float2 *iptr);
-float3 __ovld modf(float3 x, float3 *iptr);
-float4 __ovld modf(float4 x, float4 *iptr);
-float8 __ovld modf(float8 x, float8 *iptr);
-float16 __ovld modf(float16 x, float16 *iptr);
-#ifdef cl_khr_fp64
-double __ovld modf(double x, double *iptr);
-double2 __ovld modf(double2 x, double2 *iptr);
-double3 __ovld modf(double3 x, double3 *iptr);
-double4 __ovld modf(double4 x, double4 *iptr);
-double8 __ovld modf(double8 x, double8 *iptr);
-double16 __ovld modf(double16 x, double16 *iptr);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld modf(half x, half *iptr);
-half2 __ovld modf(half2 x, half2 *iptr);
-half3 __ovld modf(half3 x, half3 *iptr);
-half4 __ovld modf(half4 x, half4 *iptr);
-half8 __ovld modf(half8 x, half8 *iptr);
-half16 __ovld modf(half16 x, half16 *iptr);
-#endif //cl_khr_fp16
-#else
-float __ovld modf(float x, __global float *iptr);
-float2 __ovld modf(float2 x, __global float2 *iptr);
-float3 __ovld modf(float3 x, __global float3 *iptr);
-float4 __ovld modf(float4 x, __global float4 *iptr);
-float8 __ovld modf(float8 x, __global float8 *iptr);
-float16 __ovld modf(float16 x, __global float16 *iptr);
-float __ovld modf(float x, __local float *iptr);
-float2 __ovld modf(float2 x, __local float2 *iptr);
-float3 __ovld modf(float3 x, __local float3 *iptr);
-float4 __ovld modf(float4 x, __local float4 *iptr);
-float8 __ovld modf(float8 x, __local float8 *iptr);
-float16 __ovld modf(float16 x, __local float16 *iptr);
-float __ovld modf(float x, __private float *iptr);
-float2 __ovld modf(float2 x, __private float2 *iptr);
-float3 __ovld modf(float3 x, __private float3 *iptr);
-float4 __ovld modf(float4 x, __private float4 *iptr);
-float8 __ovld modf(float8 x, __private float8 *iptr);
-float16 __ovld modf(float16 x, __private float16 *iptr);
-#ifdef cl_khr_fp64
-double __ovld modf(double x, __global double *iptr);
-double2 __ovld modf(double2 x, __global double2 *iptr);
-double3 __ovld modf(double3 x, __global double3 *iptr);
-double4 __ovld modf(double4 x, __global double4 *iptr);
-double8 __ovld modf(double8 x, __global double8 *iptr);
-double16 __ovld modf(double16 x, __global double16 *iptr);
-double __ovld modf(double x, __local double *iptr);
-double2 __ovld modf(double2 x, __local double2 *iptr);
-double3 __ovld modf(double3 x, __local double3 *iptr);
-double4 __ovld modf(double4 x, __local double4 *iptr);
-double8 __ovld modf(double8 x, __local double8 *iptr);
-double16 __ovld modf(double16 x, __local double16 *iptr);
-double __ovld modf(double x, __private double *iptr);
-double2 __ovld modf(double2 x, __private double2 *iptr);
-double3 __ovld modf(double3 x, __private double3 *iptr);
-double4 __ovld modf(double4 x, __private double4 *iptr);
-double8 __ovld modf(double8 x, __private double8 *iptr);
-double16 __ovld modf(double16 x, __private double16 *iptr);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld modf(half x, __global half *iptr);
-half2 __ovld modf(half2 x, __global half2 *iptr);
-half3 __ovld modf(half3 x, __global half3 *iptr);
-half4 __ovld modf(half4 x, __global half4 *iptr);
-half8 __ovld modf(half8 x, __global half8 *iptr);
-half16 __ovld modf(half16 x, __global half16 *iptr);
-half __ovld modf(half x, __local half *iptr);
-half2 __ovld modf(half2 x, __local half2 *iptr);
-half3 __ovld modf(half3 x, __local half3 *iptr);
-half4 __ovld modf(half4 x, __local half4 *iptr);
-half8 __ovld modf(half8 x, __local half8 *iptr);
-half16 __ovld modf(half16 x, __local half16 *iptr);
-half __ovld modf(half x, __private half *iptr);
-half2 __ovld modf(half2 x, __private half2 *iptr);
-half3 __ovld modf(half3 x, __private half3 *iptr);
-half4 __ovld modf(half4 x, __private half4 *iptr);
-half8 __ovld modf(half8 x, __private half8 *iptr);
-half16 __ovld modf(half16 x, __private half16 *iptr);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Returns a quiet NaN. The nancode may be placed
- * in the significand of the resulting NaN.
- */
-float __ovld __cnfn nan(uint nancode);
-float2 __ovld __cnfn nan(uint2 nancode);
-float3 __ovld __cnfn nan(uint3 nancode);
-float4 __ovld __cnfn nan(uint4 nancode);
-float8 __ovld __cnfn nan(uint8 nancode);
-float16 __ovld __cnfn nan(uint16 nancode);
-#ifdef cl_khr_fp64
-double __ovld __cnfn nan(ulong nancode);
-double2 __ovld __cnfn nan(ulong2 nancode);
-double3 __ovld __cnfn nan(ulong3 nancode);
-double4 __ovld __cnfn nan(ulong4 nancode);
-double8 __ovld __cnfn nan(ulong8 nancode);
-double16 __ovld __cnfn nan(ulong16 nancode);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn nan(ushort nancode);
-half2 __ovld __cnfn nan(ushort2 nancode);
-half3 __ovld __cnfn nan(ushort3 nancode);
-half4 __ovld __cnfn nan(ushort4 nancode);
-half8 __ovld __cnfn nan(ushort8 nancode);
-half16 __ovld __cnfn nan(ushort16 nancode);
-#endif //cl_khr_fp16
-
-/**
- * Computes the next representable single-precision
- * floating-point value following x in the direction of
- * y. Thus, if y is less than x, nextafter() returns the
- * largest representable floating-point number less
- * than x.
- */
-float __ovld __cnfn nextafter(float x, float y);
-float2 __ovld __cnfn nextafter(float2 x, float2 y);
-float3 __ovld __cnfn nextafter(float3 x, float3 y);
-float4 __ovld __cnfn nextafter(float4 x, float4 y);
-float8 __ovld __cnfn nextafter(float8 x, float8 y);
-float16 __ovld __cnfn nextafter(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn nextafter(double x, double y);
-double2 __ovld __cnfn nextafter(double2 x, double2 y);
-double3 __ovld __cnfn nextafter(double3 x, double3 y);
-double4 __ovld __cnfn nextafter(double4 x, double4 y);
-double8 __ovld __cnfn nextafter(double8 x, double8 y);
-double16 __ovld __cnfn nextafter(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn nextafter(half x, half y);
-half2 __ovld __cnfn nextafter(half2 x, half2 y);
-half3 __ovld __cnfn nextafter(half3 x, half3 y);
-half4 __ovld __cnfn nextafter(half4 x, half4 y);
-half8 __ovld __cnfn nextafter(half8 x, half8 y);
-half16 __ovld __cnfn nextafter(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Compute x to the power y.
- */
-float __ovld __cnfn pow(float x, float y);
-float2 __ovld __cnfn pow(float2 x, float2 y);
-float3 __ovld __cnfn pow(float3 x, float3 y);
-float4 __ovld __cnfn pow(float4 x, float4 y);
-float8 __ovld __cnfn pow(float8 x, float8 y);
-float16 __ovld __cnfn pow(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn pow(double x, double y);
-double2 __ovld __cnfn pow(double2 x, double2 y);
-double3 __ovld __cnfn pow(double3 x, double3 y);
-double4 __ovld __cnfn pow(double4 x, double4 y);
-double8 __ovld __cnfn pow(double8 x, double8 y);
-double16 __ovld __cnfn pow(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn pow(half x, half y);
-half2 __ovld __cnfn pow(half2 x, half2 y);
-half3 __ovld __cnfn pow(half3 x, half3 y);
-half4 __ovld __cnfn pow(half4 x, half4 y);
-half8 __ovld __cnfn pow(half8 x, half8 y);
-half16 __ovld __cnfn pow(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Compute x to the power y, where y is an integer.
- */
-float __ovld __cnfn pown(float x, int y);
-float2 __ovld __cnfn pown(float2 x, int2 y);
-float3 __ovld __cnfn pown(float3 x, int3 y);
-float4 __ovld __cnfn pown(float4 x, int4 y);
-float8 __ovld __cnfn pown(float8 x, int8 y);
-float16 __ovld __cnfn pown(float16 x, int16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn pown(double x, int y);
-double2 __ovld __cnfn pown(double2 x, int2 y);
-double3 __ovld __cnfn pown(double3 x, int3 y);
-double4 __ovld __cnfn pown(double4 x, int4 y);
-double8 __ovld __cnfn pown(double8 x, int8 y);
-double16 __ovld __cnfn pown(double16 x, int16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn pown(half x, int y);
-half2 __ovld __cnfn pown(half2 x, int2 y);
-half3 __ovld __cnfn pown(half3 x, int3 y);
-half4 __ovld __cnfn pown(half4 x, int4 y);
-half8 __ovld __cnfn pown(half8 x, int8 y);
-half16 __ovld __cnfn pown(half16 x, int16 y);
-#endif //cl_khr_fp16
-
-/**
- * Compute x to the power y, where x is >= 0.
- */
-float __ovld __cnfn powr(float x, float y);
-float2 __ovld __cnfn powr(float2 x, float2 y);
-float3 __ovld __cnfn powr(float3 x, float3 y);
-float4 __ovld __cnfn powr(float4 x, float4 y);
-float8 __ovld __cnfn powr(float8 x, float8 y);
-float16 __ovld __cnfn powr(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn powr(double x, double y);
-double2 __ovld __cnfn powr(double2 x, double2 y);
-double3 __ovld __cnfn powr(double3 x, double3 y);
-double4 __ovld __cnfn powr(double4 x, double4 y);
-double8 __ovld __cnfn powr(double8 x, double8 y);
-double16 __ovld __cnfn powr(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn powr(half x, half y);
-half2 __ovld __cnfn powr(half2 x, half2 y);
-half3 __ovld __cnfn powr(half3 x, half3 y);
-half4 __ovld __cnfn powr(half4 x, half4 y);
-half8 __ovld __cnfn powr(half8 x, half8 y);
-half16 __ovld __cnfn powr(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Compute the value r such that r = x - n*y, where n
- * is the integer nearest the exact value of x/y. If there
- * are two integers closest to x/y, n shall be the even
- * one. If r is zero, it is given the same sign as x.
- */
-float __ovld __cnfn remainder(float x, float y);
-float2 __ovld __cnfn remainder(float2 x, float2 y);
-float3 __ovld __cnfn remainder(float3 x, float3 y);
-float4 __ovld __cnfn remainder(float4 x, float4 y);
-float8 __ovld __cnfn remainder(float8 x, float8 y);
-float16 __ovld __cnfn remainder(float16 x, float16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn remainder(double x, double y);
-double2 __ovld __cnfn remainder(double2 x, double2 y);
-double3 __ovld __cnfn remainder(double3 x, double3 y);
-double4 __ovld __cnfn remainder(double4 x, double4 y);
-double8 __ovld __cnfn remainder(double8 x, double8 y);
-double16 __ovld __cnfn remainder(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn remainder(half x, half y);
-half2 __ovld __cnfn remainder(half2 x, half2 y);
-half3 __ovld __cnfn remainder(half3 x, half3 y);
-half4 __ovld __cnfn remainder(half4 x, half4 y);
-half8 __ovld __cnfn remainder(half8 x, half8 y);
-half16 __ovld __cnfn remainder(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * The remquo function computes the value r such
- * that r = x - n*y, where n is the integer nearest the
- * exact value of x/y. If there are two integers closest
- * to x/y, n shall be the even one. If r is zero, it is
- * given the same sign as x. This is the same value
- * that is returned by the remainder function.
- * remquo also calculates the lower seven bits of the
- * integral quotient x/y, and gives that value the same
- * sign as x/y. It stores this signed value in the object
- * pointed to by quo.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld remquo(float x, float y, int *quo);
-float2 __ovld remquo(float2 x, float2 y, int2 *quo);
-float3 __ovld remquo(float3 x, float3 y, int3 *quo);
-float4 __ovld remquo(float4 x, float4 y, int4 *quo);
-float8 __ovld remquo(float8 x, float8 y, int8 *quo);
-float16 __ovld remquo(float16 x, float16 y, int16 *quo);
-#ifdef cl_khr_fp64
-double __ovld remquo(double x, double y, int *quo);
-double2 __ovld remquo(double2 x, double2 y, int2 *quo);
-double3 __ovld remquo(double3 x, double3 y, int3 *quo);
-double4 __ovld remquo(double4 x, double4 y, int4 *quo);
-double8 __ovld remquo(double8 x, double8 y, int8 *quo);
-double16 __ovld remquo(double16 x, double16 y, int16 *quo);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld remquo(half x, half y, int *quo);
-half2 __ovld remquo(half2 x, half2 y, int2 *quo);
-half3 __ovld remquo(half3 x, half3 y, int3 *quo);
-half4 __ovld remquo(half4 x, half4 y, int4 *quo);
-half8 __ovld remquo(half8 x, half8 y, int8 *quo);
-half16 __ovld remquo(half16 x, half16 y, int16 *quo);
-
-#endif //cl_khr_fp16
-#else
-float __ovld remquo(float x, float y, __global int *quo);
-float2 __ovld remquo(float2 x, float2 y, __global int2 *quo);
-float3 __ovld remquo(float3 x, float3 y, __global int3 *quo);
-float4 __ovld remquo(float4 x, float4 y, __global int4 *quo);
-float8 __ovld remquo(float8 x, float8 y, __global int8 *quo);
-float16 __ovld remquo(float16 x, float16 y, __global int16 *quo);
-float __ovld remquo(float x, float y, __local int *quo);
-float2 __ovld remquo(float2 x, float2 y, __local int2 *quo);
-float3 __ovld remquo(float3 x, float3 y, __local int3 *quo);
-float4 __ovld remquo(float4 x, float4 y, __local int4 *quo);
-float8 __ovld remquo(float8 x, float8 y, __local int8 *quo);
-float16 __ovld remquo(float16 x, float16 y, __local int16 *quo);
-float __ovld remquo(float x, float y, __private int *quo);
-float2 __ovld remquo(float2 x, float2 y, __private int2 *quo);
-float3 __ovld remquo(float3 x, float3 y, __private int3 *quo);
-float4 __ovld remquo(float4 x, float4 y, __private int4 *quo);
-float8 __ovld remquo(float8 x, float8 y, __private int8 *quo);
-float16 __ovld remquo(float16 x, float16 y, __private int16 *quo);
-#ifdef cl_khr_fp64
-double __ovld remquo(double x, double y, __global int *quo);
-double2 __ovld remquo(double2 x, double2 y, __global int2 *quo);
-double3 __ovld remquo(double3 x, double3 y, __global int3 *quo);
-double4 __ovld remquo(double4 x, double4 y, __global int4 *quo);
-double8 __ovld remquo(double8 x, double8 y, __global int8 *quo);
-double16 __ovld remquo(double16 x, double16 y, __global int16 *quo);
-double __ovld remquo(double x, double y, __local int *quo);
-double2 __ovld remquo(double2 x, double2 y, __local int2 *quo);
-double3 __ovld remquo(double3 x, double3 y, __local int3 *quo);
-double4 __ovld remquo(double4 x, double4 y, __local int4 *quo);
-double8 __ovld remquo(double8 x, double8 y, __local int8 *quo);
-double16 __ovld remquo(double16 x, double16 y, __local int16 *quo);
-double __ovld remquo(double x, double y, __private int *quo);
-double2 __ovld remquo(double2 x, double2 y, __private int2 *quo);
-double3 __ovld remquo(double3 x, double3 y, __private int3 *quo);
-double4 __ovld remquo(double4 x, double4 y, __private int4 *quo);
-double8 __ovld remquo(double8 x, double8 y, __private int8 *quo);
-double16 __ovld remquo(double16 x, double16 y, __private int16 *quo);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld remquo(half x, half y, __global int *quo);
-half2 __ovld remquo(half2 x, half2 y, __global int2 *quo);
-half3 __ovld remquo(half3 x, half3 y, __global int3 *quo);
-half4 __ovld remquo(half4 x, half4 y, __global int4 *quo);
-half8 __ovld remquo(half8 x, half8 y, __global int8 *quo);
-half16 __ovld remquo(half16 x, half16 y, __global int16 *quo);
-half __ovld remquo(half x, half y, __local int *quo);
-half2 __ovld remquo(half2 x, half2 y, __local int2 *quo);
-half3 __ovld remquo(half3 x, half3 y, __local int3 *quo);
-half4 __ovld remquo(half4 x, half4 y, __local int4 *quo);
-half8 __ovld remquo(half8 x, half8 y, __local int8 *quo);
-half16 __ovld remquo(half16 x, half16 y, __local int16 *quo);
-half __ovld remquo(half x, half y, __private int *quo);
-half2 __ovld remquo(half2 x, half2 y, __private int2 *quo);
-half3 __ovld remquo(half3 x, half3 y, __private int3 *quo);
-half4 __ovld remquo(half4 x, half4 y, __private int4 *quo);
-half8 __ovld remquo(half8 x, half8 y, __private int8 *quo);
-half16 __ovld remquo(half16 x, half16 y, __private int16 *quo);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-/**
- * Round to integral value (using round to nearest
- * even rounding mode) in floating-point format.
- * Refer to section 7.1 for description of rounding
- * modes.
- */
-float __ovld __cnfn rint(float);
-float2 __ovld __cnfn rint(float2);
-float3 __ovld __cnfn rint(float3);
-float4 __ovld __cnfn rint(float4);
-float8 __ovld __cnfn rint(float8);
-float16 __ovld __cnfn rint(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn rint(double);
-double2 __ovld __cnfn rint(double2);
-double3 __ovld __cnfn rint(double3);
-double4 __ovld __cnfn rint(double4);
-double8 __ovld __cnfn rint(double8);
-double16 __ovld __cnfn rint(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn rint(half);
-half2 __ovld __cnfn rint(half2);
-half3 __ovld __cnfn rint(half3);
-half4 __ovld __cnfn rint(half4);
-half8 __ovld __cnfn rint(half8);
-half16 __ovld __cnfn rint(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute x to the power 1/y.
- */
-float __ovld __cnfn rootn(float x, int y);
-float2 __ovld __cnfn rootn(float2 x, int2 y);
-float3 __ovld __cnfn rootn(float3 x, int3 y);
-float4 __ovld __cnfn rootn(float4 x, int4 y);
-float8 __ovld __cnfn rootn(float8 x, int8 y);
-float16 __ovld __cnfn rootn(float16 x, int16 y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn rootn(double x, int y);
-double2 __ovld __cnfn rootn(double2 x, int2 y);
-double3 __ovld __cnfn rootn(double3 x, int3 y);
-double4 __ovld __cnfn rootn(double4 x, int4 y);
-double8 __ovld __cnfn rootn(double8 x, int8 y);
-double16 __ovld __cnfn rootn(double16 x, int16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn rootn(half x, int y);
-half2 __ovld __cnfn rootn(half2 x, int2 y);
-half3 __ovld __cnfn rootn(half3 x, int3 y);
-half4 __ovld __cnfn rootn(half4 x, int4 y);
-half8 __ovld __cnfn rootn(half8 x, int8 y);
-half16 __ovld __cnfn rootn(half16 x, int16 y);
-#endif //cl_khr_fp16
-
-/**
- * Return the integral value nearest to x rounding
- * halfway cases away from zero, regardless of the
- * current rounding direction.
- */
-float __ovld __cnfn round(float x);
-float2 __ovld __cnfn round(float2 x);
-float3 __ovld __cnfn round(float3 x);
-float4 __ovld __cnfn round(float4 x);
-float8 __ovld __cnfn round(float8 x);
-float16 __ovld __cnfn round(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn round(double x);
-double2 __ovld __cnfn round(double2 x);
-double3 __ovld __cnfn round(double3 x);
-double4 __ovld __cnfn round(double4 x);
-double8 __ovld __cnfn round(double8 x);
-double16 __ovld __cnfn round(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn round(half x);
-half2 __ovld __cnfn round(half2 x);
-half3 __ovld __cnfn round(half3 x);
-half4 __ovld __cnfn round(half4 x);
-half8 __ovld __cnfn round(half8 x);
-half16 __ovld __cnfn round(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute inverse square root.
- */
-float __ovld __cnfn rsqrt(float);
-float2 __ovld __cnfn rsqrt(float2);
-float3 __ovld __cnfn rsqrt(float3);
-float4 __ovld __cnfn rsqrt(float4);
-float8 __ovld __cnfn rsqrt(float8);
-float16 __ovld __cnfn rsqrt(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn rsqrt(double);
-double2 __ovld __cnfn rsqrt(double2);
-double3 __ovld __cnfn rsqrt(double3);
-double4 __ovld __cnfn rsqrt(double4);
-double8 __ovld __cnfn rsqrt(double8);
-double16 __ovld __cnfn rsqrt(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn rsqrt(half);
-half2 __ovld __cnfn rsqrt(half2);
-half3 __ovld __cnfn rsqrt(half3);
-half4 __ovld __cnfn rsqrt(half4);
-half8 __ovld __cnfn rsqrt(half8);
-half16 __ovld __cnfn rsqrt(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute sine.
- */
-float __ovld __cnfn sin(float);
-float2 __ovld __cnfn sin(float2);
-float3 __ovld __cnfn sin(float3);
-float4 __ovld __cnfn sin(float4);
-float8 __ovld __cnfn sin(float8);
-float16 __ovld __cnfn sin(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn sin(double);
-double2 __ovld __cnfn sin(double2);
-double3 __ovld __cnfn sin(double3);
-double4 __ovld __cnfn sin(double4);
-double8 __ovld __cnfn sin(double8);
-double16 __ovld __cnfn sin(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn sin(half);
-half2 __ovld __cnfn sin(half2);
-half3 __ovld __cnfn sin(half3);
-half4 __ovld __cnfn sin(half4);
-half8 __ovld __cnfn sin(half8);
-half16 __ovld __cnfn sin(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute sine and cosine of x. The computed sine
- * is the return value and computed cosine is returned
- * in cosval.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld sincos(float x, float *cosval);
-float2 __ovld sincos(float2 x, float2 *cosval);
-float3 __ovld sincos(float3 x, float3 *cosval);
-float4 __ovld sincos(float4 x, float4 *cosval);
-float8 __ovld sincos(float8 x, float8 *cosval);
-float16 __ovld sincos(float16 x, float16 *cosval);
-#ifdef cl_khr_fp64
-double __ovld sincos(double x, double *cosval);
-double2 __ovld sincos(double2 x, double2 *cosval);
-double3 __ovld sincos(double3 x, double3 *cosval);
-double4 __ovld sincos(double4 x, double4 *cosval);
-double8 __ovld sincos(double8 x, double8 *cosval);
-double16 __ovld sincos(double16 x, double16 *cosval);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld sincos(half x, half *cosval);
-half2 __ovld sincos(half2 x, half2 *cosval);
-half3 __ovld sincos(half3 x, half3 *cosval);
-half4 __ovld sincos(half4 x, half4 *cosval);
-half8 __ovld sincos(half8 x, half8 *cosval);
-half16 __ovld sincos(half16 x, half16 *cosval);
-#endif //cl_khr_fp16
-#else
-float __ovld sincos(float x, __global float *cosval);
-float2 __ovld sincos(float2 x, __global float2 *cosval);
-float3 __ovld sincos(float3 x, __global float3 *cosval);
-float4 __ovld sincos(float4 x, __global float4 *cosval);
-float8 __ovld sincos(float8 x, __global float8 *cosval);
-float16 __ovld sincos(float16 x, __global float16 *cosval);
-float __ovld sincos(float x, __local float *cosval);
-float2 __ovld sincos(float2 x, __local float2 *cosval);
-float3 __ovld sincos(float3 x, __local float3 *cosval);
-float4 __ovld sincos(float4 x, __local float4 *cosval);
-float8 __ovld sincos(float8 x, __local float8 *cosval);
-float16 __ovld sincos(float16 x, __local float16 *cosval);
-float __ovld sincos(float x, __private float *cosval);
-float2 __ovld sincos(float2 x, __private float2 *cosval);
-float3 __ovld sincos(float3 x, __private float3 *cosval);
-float4 __ovld sincos(float4 x, __private float4 *cosval);
-float8 __ovld sincos(float8 x, __private float8 *cosval);
-float16 __ovld sincos(float16 x, __private float16 *cosval);
-#ifdef cl_khr_fp64
-double __ovld sincos(double x, __global double *cosval);
-double2 __ovld sincos(double2 x, __global double2 *cosval);
-double3 __ovld sincos(double3 x, __global double3 *cosval);
-double4 __ovld sincos(double4 x, __global double4 *cosval);
-double8 __ovld sincos(double8 x, __global double8 *cosval);
-double16 __ovld sincos(double16 x, __global double16 *cosval);
-double __ovld sincos(double x, __local double *cosval);
-double2 __ovld sincos(double2 x, __local double2 *cosval);
-double3 __ovld sincos(double3 x, __local double3 *cosval);
-double4 __ovld sincos(double4 x, __local double4 *cosval);
-double8 __ovld sincos(double8 x, __local double8 *cosval);
-double16 __ovld sincos(double16 x, __local double16 *cosval);
-double __ovld sincos(double x, __private double *cosval);
-double2 __ovld sincos(double2 x, __private double2 *cosval);
-double3 __ovld sincos(double3 x, __private double3 *cosval);
-double4 __ovld sincos(double4 x, __private double4 *cosval);
-double8 __ovld sincos(double8 x, __private double8 *cosval);
-double16 __ovld sincos(double16 x, __private double16 *cosval);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld sincos(half x, __global half *cosval);
-half2 __ovld sincos(half2 x, __global half2 *cosval);
-half3 __ovld sincos(half3 x, __global half3 *cosval);
-half4 __ovld sincos(half4 x, __global half4 *cosval);
-half8 __ovld sincos(half8 x, __global half8 *cosval);
-half16 __ovld sincos(half16 x, __global half16 *cosval);
-half __ovld sincos(half x, __local half *cosval);
-half2 __ovld sincos(half2 x, __local half2 *cosval);
-half3 __ovld sincos(half3 x, __local half3 *cosval);
-half4 __ovld sincos(half4 x, __local half4 *cosval);
-half8 __ovld sincos(half8 x, __local half8 *cosval);
-half16 __ovld sincos(half16 x, __local half16 *cosval);
-half __ovld sincos(half x, __private half *cosval);
-half2 __ovld sincos(half2 x, __private half2 *cosval);
-half3 __ovld sincos(half3 x, __private half3 *cosval);
-half4 __ovld sincos(half4 x, __private half4 *cosval);
-half8 __ovld sincos(half8 x, __private half8 *cosval);
-half16 __ovld sincos(half16 x, __private half16 *cosval);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Compute hyperbolic sine.
- */
-float __ovld __cnfn sinh(float);
-float2 __ovld __cnfn sinh(float2);
-float3 __ovld __cnfn sinh(float3);
-float4 __ovld __cnfn sinh(float4);
-float8 __ovld __cnfn sinh(float8);
-float16 __ovld __cnfn sinh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn sinh(double);
-double2 __ovld __cnfn sinh(double2);
-double3 __ovld __cnfn sinh(double3);
-double4 __ovld __cnfn sinh(double4);
-double8 __ovld __cnfn sinh(double8);
-double16 __ovld __cnfn sinh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn sinh(half);
-half2 __ovld __cnfn sinh(half2);
-half3 __ovld __cnfn sinh(half3);
-half4 __ovld __cnfn sinh(half4);
-half8 __ovld __cnfn sinh(half8);
-half16 __ovld __cnfn sinh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute sin (PI * x).
- */
-float __ovld __cnfn sinpi(float x);
-float2 __ovld __cnfn sinpi(float2 x);
-float3 __ovld __cnfn sinpi(float3 x);
-float4 __ovld __cnfn sinpi(float4 x);
-float8 __ovld __cnfn sinpi(float8 x);
-float16 __ovld __cnfn sinpi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn sinpi(double x);
-double2 __ovld __cnfn sinpi(double2 x);
-double3 __ovld __cnfn sinpi(double3 x);
-double4 __ovld __cnfn sinpi(double4 x);
-double8 __ovld __cnfn sinpi(double8 x);
-double16 __ovld __cnfn sinpi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn sinpi(half x);
-half2 __ovld __cnfn sinpi(half2 x);
-half3 __ovld __cnfn sinpi(half3 x);
-half4 __ovld __cnfn sinpi(half4 x);
-half8 __ovld __cnfn sinpi(half8 x);
-half16 __ovld __cnfn sinpi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute square root.
- */
-float __ovld __cnfn sqrt(float);
-float2 __ovld __cnfn sqrt(float2);
-float3 __ovld __cnfn sqrt(float3);
-float4 __ovld __cnfn sqrt(float4);
-float8 __ovld __cnfn sqrt(float8);
-float16 __ovld __cnfn sqrt(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn sqrt(double);
-double2 __ovld __cnfn sqrt(double2);
-double3 __ovld __cnfn sqrt(double3);
-double4 __ovld __cnfn sqrt(double4);
-double8 __ovld __cnfn sqrt(double8);
-double16 __ovld __cnfn sqrt(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn sqrt(half);
-half2 __ovld __cnfn sqrt(half2);
-half3 __ovld __cnfn sqrt(half3);
-half4 __ovld __cnfn sqrt(half4);
-half8 __ovld __cnfn sqrt(half8);
-half16 __ovld __cnfn sqrt(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute tangent.
- */
-float __ovld __cnfn tan(float);
-float2 __ovld __cnfn tan(float2);
-float3 __ovld __cnfn tan(float3);
-float4 __ovld __cnfn tan(float4);
-float8 __ovld __cnfn tan(float8);
-float16 __ovld __cnfn tan(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn tan(double);
-double2 __ovld __cnfn tan(double2);
-double3 __ovld __cnfn tan(double3);
-double4 __ovld __cnfn tan(double4);
-double8 __ovld __cnfn tan(double8);
-double16 __ovld __cnfn tan(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn tan(half);
-half2 __ovld __cnfn tan(half2);
-half3 __ovld __cnfn tan(half3);
-half4 __ovld __cnfn tan(half4);
-half8 __ovld __cnfn tan(half8);
-half16 __ovld __cnfn tan(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute hyperbolic tangent.
- */
-float __ovld __cnfn tanh(float);
-float2 __ovld __cnfn tanh(float2);
-float3 __ovld __cnfn tanh(float3);
-float4 __ovld __cnfn tanh(float4);
-float8 __ovld __cnfn tanh(float8);
-float16 __ovld __cnfn tanh(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn tanh(double);
-double2 __ovld __cnfn tanh(double2);
-double3 __ovld __cnfn tanh(double3);
-double4 __ovld __cnfn tanh(double4);
-double8 __ovld __cnfn tanh(double8);
-double16 __ovld __cnfn tanh(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn tanh(half);
-half2 __ovld __cnfn tanh(half2);
-half3 __ovld __cnfn tanh(half3);
-half4 __ovld __cnfn tanh(half4);
-half8 __ovld __cnfn tanh(half8);
-half16 __ovld __cnfn tanh(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute tan (PI * x).
- */
-float __ovld __cnfn tanpi(float x);
-float2 __ovld __cnfn tanpi(float2 x);
-float3 __ovld __cnfn tanpi(float3 x);
-float4 __ovld __cnfn tanpi(float4 x);
-float8 __ovld __cnfn tanpi(float8 x);
-float16 __ovld __cnfn tanpi(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn tanpi(double x);
-double2 __ovld __cnfn tanpi(double2 x);
-double3 __ovld __cnfn tanpi(double3 x);
-double4 __ovld __cnfn tanpi(double4 x);
-double8 __ovld __cnfn tanpi(double8 x);
-double16 __ovld __cnfn tanpi(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn tanpi(half x);
-half2 __ovld __cnfn tanpi(half2 x);
-half3 __ovld __cnfn tanpi(half3 x);
-half4 __ovld __cnfn tanpi(half4 x);
-half8 __ovld __cnfn tanpi(half8 x);
-half16 __ovld __cnfn tanpi(half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Compute the gamma function.
- */
-float __ovld __cnfn tgamma(float);
-float2 __ovld __cnfn tgamma(float2);
-float3 __ovld __cnfn tgamma(float3);
-float4 __ovld __cnfn tgamma(float4);
-float8 __ovld __cnfn tgamma(float8);
-float16 __ovld __cnfn tgamma(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn tgamma(double);
-double2 __ovld __cnfn tgamma(double2);
-double3 __ovld __cnfn tgamma(double3);
-double4 __ovld __cnfn tgamma(double4);
-double8 __ovld __cnfn tgamma(double8);
-double16 __ovld __cnfn tgamma(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn tgamma(half);
-half2 __ovld __cnfn tgamma(half2);
-half3 __ovld __cnfn tgamma(half3);
-half4 __ovld __cnfn tgamma(half4);
-half8 __ovld __cnfn tgamma(half8);
-half16 __ovld __cnfn tgamma(half16);
-#endif //cl_khr_fp16
-
-/**
- * Round to integral value using the round to zero
- * rounding mode.
- */
-float __ovld __cnfn trunc(float);
-float2 __ovld __cnfn trunc(float2);
-float3 __ovld __cnfn trunc(float3);
-float4 __ovld __cnfn trunc(float4);
-float8 __ovld __cnfn trunc(float8);
-float16 __ovld __cnfn trunc(float16);
-#ifdef cl_khr_fp64
-double __ovld __cnfn trunc(double);
-double2 __ovld __cnfn trunc(double2);
-double3 __ovld __cnfn trunc(double3);
-double4 __ovld __cnfn trunc(double4);
-double8 __ovld __cnfn trunc(double8);
-double16 __ovld __cnfn trunc(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn trunc(half);
-half2 __ovld __cnfn trunc(half2);
-half3 __ovld __cnfn trunc(half3);
-half4 __ovld __cnfn trunc(half4);
-half8 __ovld __cnfn trunc(half8);
-half16 __ovld __cnfn trunc(half16);
-#endif //cl_khr_fp16
-
-/**
- * Compute cosine. x must be in the range -2^16 ... +2^16.
- */
-float __ovld __cnfn half_cos(float x);
-float2 __ovld __cnfn half_cos(float2 x);
-float3 __ovld __cnfn half_cos(float3 x);
-float4 __ovld __cnfn half_cos(float4 x);
-float8 __ovld __cnfn half_cos(float8 x);
-float16 __ovld __cnfn half_cos(float16 x);
-
-/**
- * Compute x / y.
- */
-float __ovld __cnfn half_divide(float x, float y);
-float2 __ovld __cnfn half_divide(float2 x, float2 y);
-float3 __ovld __cnfn half_divide(float3 x, float3 y);
-float4 __ovld __cnfn half_divide(float4 x, float4 y);
-float8 __ovld __cnfn half_divide(float8 x, float8 y);
-float16 __ovld __cnfn half_divide(float16 x, float16 y);
-
-/**
- * Compute the base- e exponential of x.
- */
-float __ovld __cnfn half_exp(float x);
-float2 __ovld __cnfn half_exp(float2 x);
-float3 __ovld __cnfn half_exp(float3 x);
-float4 __ovld __cnfn half_exp(float4 x);
-float8 __ovld __cnfn half_exp(float8 x);
-float16 __ovld __cnfn half_exp(float16 x);
-
-/**
- * Compute the base- 2 exponential of x.
- */
-float __ovld __cnfn half_exp2(float x);
-float2 __ovld __cnfn half_exp2(float2 x);
-float3 __ovld __cnfn half_exp2(float3 x);
-float4 __ovld __cnfn half_exp2(float4 x);
-float8 __ovld __cnfn half_exp2(float8 x);
-float16 __ovld __cnfn half_exp2(float16 x);
-
-/**
- * Compute the base- 10 exponential of x.
- */
-float __ovld __cnfn half_exp10(float x);
-float2 __ovld __cnfn half_exp10(float2 x);
-float3 __ovld __cnfn half_exp10(float3 x);
-float4 __ovld __cnfn half_exp10(float4 x);
-float8 __ovld __cnfn half_exp10(float8 x);
-float16 __ovld __cnfn half_exp10(float16 x);
-
-/**
- * Compute natural logarithm.
- */
-float __ovld __cnfn half_log(float x);
-float2 __ovld __cnfn half_log(float2 x);
-float3 __ovld __cnfn half_log(float3 x);
-float4 __ovld __cnfn half_log(float4 x);
-float8 __ovld __cnfn half_log(float8 x);
-float16 __ovld __cnfn half_log(float16 x);
-
-/**
- * Compute a base 2 logarithm.
- */
-float __ovld __cnfn half_log2(float x);
-float2 __ovld __cnfn half_log2(float2 x);
-float3 __ovld __cnfn half_log2(float3 x);
-float4 __ovld __cnfn half_log2(float4 x);
-float8 __ovld __cnfn half_log2(float8 x);
-float16 __ovld __cnfn half_log2(float16 x);
-
-/**
- * Compute a base 10 logarithm.
- */
-float __ovld __cnfn half_log10(float x);
-float2 __ovld __cnfn half_log10(float2 x);
-float3 __ovld __cnfn half_log10(float3 x);
-float4 __ovld __cnfn half_log10(float4 x);
-float8 __ovld __cnfn half_log10(float8 x);
-float16 __ovld __cnfn half_log10(float16 x);
-
-/**
- * Compute x to the power y, where x is >= 0.
- */
-float __ovld __cnfn half_powr(float x, float y);
-float2 __ovld __cnfn half_powr(float2 x, float2 y);
-float3 __ovld __cnfn half_powr(float3 x, float3 y);
-float4 __ovld __cnfn half_powr(float4 x, float4 y);
-float8 __ovld __cnfn half_powr(float8 x, float8 y);
-float16 __ovld __cnfn half_powr(float16 x, float16 y);
-
-/**
- * Compute reciprocal.
- */
-float __ovld __cnfn half_recip(float x);
-float2 __ovld __cnfn half_recip(float2 x);
-float3 __ovld __cnfn half_recip(float3 x);
-float4 __ovld __cnfn half_recip(float4 x);
-float8 __ovld __cnfn half_recip(float8 x);
-float16 __ovld __cnfn half_recip(float16 x);
-
-/**
- * Compute inverse square root.
- */
-float __ovld __cnfn half_rsqrt(float x);
-float2 __ovld __cnfn half_rsqrt(float2 x);
-float3 __ovld __cnfn half_rsqrt(float3 x);
-float4 __ovld __cnfn half_rsqrt(float4 x);
-float8 __ovld __cnfn half_rsqrt(float8 x);
-float16 __ovld __cnfn half_rsqrt(float16 x);
-
-/**
- * Compute sine. x must be in the range -2^16 ... +2^16.
- */
-float __ovld __cnfn half_sin(float x);
-float2 __ovld __cnfn half_sin(float2 x);
-float3 __ovld __cnfn half_sin(float3 x);
-float4 __ovld __cnfn half_sin(float4 x);
-float8 __ovld __cnfn half_sin(float8 x);
-float16 __ovld __cnfn half_sin(float16 x);
-
-/**
- * Compute square root.
- */
-float __ovld __cnfn half_sqrt(float x);
-float2 __ovld __cnfn half_sqrt(float2 x);
-float3 __ovld __cnfn half_sqrt(float3 x);
-float4 __ovld __cnfn half_sqrt(float4 x);
-float8 __ovld __cnfn half_sqrt(float8 x);
-float16 __ovld __cnfn half_sqrt(float16 x);
-
-/**
- * Compute tangent. x must be in the range -216 ... +216.
- */
-float __ovld __cnfn half_tan(float x);
-float2 __ovld __cnfn half_tan(float2 x);
-float3 __ovld __cnfn half_tan(float3 x);
-float4 __ovld __cnfn half_tan(float4 x);
-float8 __ovld __cnfn half_tan(float8 x);
-float16 __ovld __cnfn half_tan(float16 x);
-
-/**
- * Compute cosine over an implementation-defined range.
- * The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_cos(float x);
-float2 __ovld __cnfn native_cos(float2 x);
-float3 __ovld __cnfn native_cos(float3 x);
-float4 __ovld __cnfn native_cos(float4 x);
-float8 __ovld __cnfn native_cos(float8 x);
-float16 __ovld __cnfn native_cos(float16 x);
-
-/**
- * Compute x / y over an implementation-defined range.
- * The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_divide(float x, float y);
-float2 __ovld __cnfn native_divide(float2 x, float2 y);
-float3 __ovld __cnfn native_divide(float3 x, float3 y);
-float4 __ovld __cnfn native_divide(float4 x, float4 y);
-float8 __ovld __cnfn native_divide(float8 x, float8 y);
-float16 __ovld __cnfn native_divide(float16 x, float16 y);
-
-/**
- * Compute the base- e exponential of x over an
- * implementation-defined range. The maximum error is
- * implementation-defined.
- */
-float __ovld __cnfn native_exp(float x);
-float2 __ovld __cnfn native_exp(float2 x);
-float3 __ovld __cnfn native_exp(float3 x);
-float4 __ovld __cnfn native_exp(float4 x);
-float8 __ovld __cnfn native_exp(float8 x);
-float16 __ovld __cnfn native_exp(float16 x);
-
-/**
- * Compute the base- 2 exponential of x over an
- * implementation-defined range. The maximum error is
- * implementation-defined.
- */
-float __ovld __cnfn native_exp2(float x);
-float2 __ovld __cnfn native_exp2(float2 x);
-float3 __ovld __cnfn native_exp2(float3 x);
-float4 __ovld __cnfn native_exp2(float4 x);
-float8 __ovld __cnfn native_exp2(float8 x);
-float16 __ovld __cnfn native_exp2(float16 x);
-
-/**
- * Compute the base- 10 exponential of x over an
- * implementation-defined range. The maximum error is
- * implementation-defined.
- */
-float __ovld __cnfn native_exp10(float x);
-float2 __ovld __cnfn native_exp10(float2 x);
-float3 __ovld __cnfn native_exp10(float3 x);
-float4 __ovld __cnfn native_exp10(float4 x);
-float8 __ovld __cnfn native_exp10(float8 x);
-float16 __ovld __cnfn native_exp10(float16 x);
-
-/**
- * Compute natural logarithm over an implementationdefined
- * range. The maximum error is implementation
- * defined.
- */
-float __ovld __cnfn native_log(float x);
-float2 __ovld __cnfn native_log(float2 x);
-float3 __ovld __cnfn native_log(float3 x);
-float4 __ovld __cnfn native_log(float4 x);
-float8 __ovld __cnfn native_log(float8 x);
-float16 __ovld __cnfn native_log(float16 x);
-
-/**
- * Compute a base 2 logarithm over an implementationdefined
- * range. The maximum error is implementationdefined.
- */
-float __ovld __cnfn native_log2(float x);
-float2 __ovld __cnfn native_log2(float2 x);
-float3 __ovld __cnfn native_log2(float3 x);
-float4 __ovld __cnfn native_log2(float4 x);
-float8 __ovld __cnfn native_log2(float8 x);
-float16 __ovld __cnfn native_log2(float16 x);
-
-/**
- * Compute a base 10 logarithm over an implementationdefined
- * range. The maximum error is implementationdefined.
- */
-float __ovld __cnfn native_log10(float x);
-float2 __ovld __cnfn native_log10(float2 x);
-float3 __ovld __cnfn native_log10(float3 x);
-float4 __ovld __cnfn native_log10(float4 x);
-float8 __ovld __cnfn native_log10(float8 x);
-float16 __ovld __cnfn native_log10(float16 x);
-
-/**
- * Compute x to the power y, where x is >= 0. The range of
- * x and y are implementation-defined. The maximum error
- * is implementation-defined.
- */
-float __ovld __cnfn native_powr(float x, float y);
-float2 __ovld __cnfn native_powr(float2 x, float2 y);
-float3 __ovld __cnfn native_powr(float3 x, float3 y);
-float4 __ovld __cnfn native_powr(float4 x, float4 y);
-float8 __ovld __cnfn native_powr(float8 x, float8 y);
-float16 __ovld __cnfn native_powr(float16 x, float16 y);
-
-/**
- * Compute reciprocal over an implementation-defined
- * range. The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_recip(float x);
-float2 __ovld __cnfn native_recip(float2 x);
-float3 __ovld __cnfn native_recip(float3 x);
-float4 __ovld __cnfn native_recip(float4 x);
-float8 __ovld __cnfn native_recip(float8 x);
-float16 __ovld __cnfn native_recip(float16 x);
-
-/**
- * Compute inverse square root over an implementationdefined
- * range. The maximum error is implementationdefined.
- */
-float __ovld __cnfn native_rsqrt(float x);
-float2 __ovld __cnfn native_rsqrt(float2 x);
-float3 __ovld __cnfn native_rsqrt(float3 x);
-float4 __ovld __cnfn native_rsqrt(float4 x);
-float8 __ovld __cnfn native_rsqrt(float8 x);
-float16 __ovld __cnfn native_rsqrt(float16 x);
-
-/**
- * Compute sine over an implementation-defined range.
- * The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_sin(float x);
-float2 __ovld __cnfn native_sin(float2 x);
-float3 __ovld __cnfn native_sin(float3 x);
-float4 __ovld __cnfn native_sin(float4 x);
-float8 __ovld __cnfn native_sin(float8 x);
-float16 __ovld __cnfn native_sin(float16 x);
-
-/**
- * Compute square root over an implementation-defined
- * range. The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_sqrt(float x);
-float2 __ovld __cnfn native_sqrt(float2 x);
-float3 __ovld __cnfn native_sqrt(float3 x);
-float4 __ovld __cnfn native_sqrt(float4 x);
-float8 __ovld __cnfn native_sqrt(float8 x);
-float16 __ovld __cnfn native_sqrt(float16 x);
-
-/**
- * Compute tangent over an implementation-defined range.
- * The maximum error is implementation-defined.
- */
-float __ovld __cnfn native_tan(float x);
-float2 __ovld __cnfn native_tan(float2 x);
-float3 __ovld __cnfn native_tan(float3 x);
-float4 __ovld __cnfn native_tan(float4 x);
-float8 __ovld __cnfn native_tan(float8 x);
-float16 __ovld __cnfn native_tan(float16 x);
-
-// OpenCL v1.1 s6.11.3, v1.2 s6.12.3, v2.0 s6.13.3 - Integer Functions
-
-/**
- * Returns | x |.
- */
-uchar __ovld __cnfn abs(char x);
-uchar __ovld __cnfn abs(uchar x);
-uchar2 __ovld __cnfn abs(char2 x);
-uchar2 __ovld __cnfn abs(uchar2 x);
-uchar3 __ovld __cnfn abs(char3 x);
-uchar3 __ovld __cnfn abs(uchar3 x);
-uchar4 __ovld __cnfn abs(char4 x);
-uchar4 __ovld __cnfn abs(uchar4 x);
-uchar8 __ovld __cnfn abs(char8 x);
-uchar8 __ovld __cnfn abs(uchar8 x);
-uchar16 __ovld __cnfn abs(char16 x);
-uchar16 __ovld __cnfn abs(uchar16 x);
-ushort __ovld __cnfn abs(short x);
-ushort __ovld __cnfn abs(ushort x);
-ushort2 __ovld __cnfn abs(short2 x);
-ushort2 __ovld __cnfn abs(ushort2 x);
-ushort3 __ovld __cnfn abs(short3 x);
-ushort3 __ovld __cnfn abs(ushort3 x);
-ushort4 __ovld __cnfn abs(short4 x);
-ushort4 __ovld __cnfn abs(ushort4 x);
-ushort8 __ovld __cnfn abs(short8 x);
-ushort8 __ovld __cnfn abs(ushort8 x);
-ushort16 __ovld __cnfn abs(short16 x);
-ushort16 __ovld __cnfn abs(ushort16 x);
-uint __ovld __cnfn abs(int x);
-uint __ovld __cnfn abs(uint x);
-uint2 __ovld __cnfn abs(int2 x);
-uint2 __ovld __cnfn abs(uint2 x);
-uint3 __ovld __cnfn abs(int3 x);
-uint3 __ovld __cnfn abs(uint3 x);
-uint4 __ovld __cnfn abs(int4 x);
-uint4 __ovld __cnfn abs(uint4 x);
-uint8 __ovld __cnfn abs(int8 x);
-uint8 __ovld __cnfn abs(uint8 x);
-uint16 __ovld __cnfn abs(int16 x);
-uint16 __ovld __cnfn abs(uint16 x);
-ulong __ovld __cnfn abs(long x);
-ulong __ovld __cnfn abs(ulong x);
-ulong2 __ovld __cnfn abs(long2 x);
-ulong2 __ovld __cnfn abs(ulong2 x);
-ulong3 __ovld __cnfn abs(long3 x);
-ulong3 __ovld __cnfn abs(ulong3 x);
-ulong4 __ovld __cnfn abs(long4 x);
-ulong4 __ovld __cnfn abs(ulong4 x);
-ulong8 __ovld __cnfn abs(long8 x);
-ulong8 __ovld __cnfn abs(ulong8 x);
-ulong16 __ovld __cnfn abs(long16 x);
-ulong16 __ovld __cnfn abs(ulong16 x);
-
-/**
- * Returns | x - y | without modulo overflow.
- */
-uchar __ovld __cnfn abs_diff(char x, char y);
-uchar __ovld __cnfn abs_diff(uchar x, uchar y);
-uchar2 __ovld __cnfn abs_diff(char2 x, char2 y);
-uchar2 __ovld __cnfn abs_diff(uchar2 x, uchar2 y);
-uchar3 __ovld __cnfn abs_diff(char3 x, char3 y);
-uchar3 __ovld __cnfn abs_diff(uchar3 x, uchar3 y);
-uchar4 __ovld __cnfn abs_diff(char4 x, char4 y);
-uchar4 __ovld __cnfn abs_diff(uchar4 x, uchar4 y);
-uchar8 __ovld __cnfn abs_diff(char8 x, char8 y);
-uchar8 __ovld __cnfn abs_diff(uchar8 x, uchar8 y);
-uchar16 __ovld __cnfn abs_diff(char16 x, char16 y);
-uchar16 __ovld __cnfn abs_diff(uchar16 x, uchar16 y);
-ushort __ovld __cnfn abs_diff(short x, short y);
-ushort __ovld __cnfn abs_diff(ushort x, ushort y);
-ushort2 __ovld __cnfn abs_diff(short2 x, short2 y);
-ushort2 __ovld __cnfn abs_diff(ushort2 x, ushort2 y);
-ushort3 __ovld __cnfn abs_diff(short3 x, short3 y);
-ushort3 __ovld __cnfn abs_diff(ushort3 x, ushort3 y);
-ushort4 __ovld __cnfn abs_diff(short4 x, short4 y);
-ushort4 __ovld __cnfn abs_diff(ushort4 x, ushort4 y);
-ushort8 __ovld __cnfn abs_diff(short8 x, short8 y);
-ushort8 __ovld __cnfn abs_diff(ushort8 x, ushort8 y);
-ushort16 __ovld __cnfn abs_diff(short16 x, short16 y);
-ushort16 __ovld __cnfn abs_diff(ushort16 x, ushort16 y);
-uint __ovld __cnfn abs_diff(int x, int y);
-uint __ovld __cnfn abs_diff(uint x, uint y);
-uint2 __ovld __cnfn abs_diff(int2 x, int2 y);
-uint2 __ovld __cnfn abs_diff(uint2 x, uint2 y);
-uint3 __ovld __cnfn abs_diff(int3 x, int3 y);
-uint3 __ovld __cnfn abs_diff(uint3 x, uint3 y);
-uint4 __ovld __cnfn abs_diff(int4 x, int4 y);
-uint4 __ovld __cnfn abs_diff(uint4 x, uint4 y);
-uint8 __ovld __cnfn abs_diff(int8 x, int8 y);
-uint8 __ovld __cnfn abs_diff(uint8 x, uint8 y);
-uint16 __ovld __cnfn abs_diff(int16 x, int16 y);
-uint16 __ovld __cnfn abs_diff(uint16 x, uint16 y);
-ulong __ovld __cnfn abs_diff(long x, long y);
-ulong __ovld __cnfn abs_diff(ulong x, ulong y);
-ulong2 __ovld __cnfn abs_diff(long2 x, long2 y);
-ulong2 __ovld __cnfn abs_diff(ulong2 x, ulong2 y);
-ulong3 __ovld __cnfn abs_diff(long3 x, long3 y);
-ulong3 __ovld __cnfn abs_diff(ulong3 x, ulong3 y);
-ulong4 __ovld __cnfn abs_diff(long4 x, long4 y);
-ulong4 __ovld __cnfn abs_diff(ulong4 x, ulong4 y);
-ulong8 __ovld __cnfn abs_diff(long8 x, long8 y);
-ulong8 __ovld __cnfn abs_diff(ulong8 x, ulong8 y);
-ulong16 __ovld __cnfn abs_diff(long16 x, long16 y);
-ulong16 __ovld __cnfn abs_diff(ulong16 x, ulong16 y);
-
-/**
- * Returns x + y and saturates the result.
- */
-char __ovld __cnfn add_sat(char x, char y);
-uchar __ovld __cnfn add_sat(uchar x, uchar y);
-char2 __ovld __cnfn add_sat(char2 x, char2 y);
-uchar2 __ovld __cnfn add_sat(uchar2 x, uchar2 y);
-char3 __ovld __cnfn add_sat(char3 x, char3 y);
-uchar3 __ovld __cnfn add_sat(uchar3 x, uchar3 y);
-char4 __ovld __cnfn add_sat(char4 x, char4 y);
-uchar4 __ovld __cnfn add_sat(uchar4 x, uchar4 y);
-char8 __ovld __cnfn add_sat(char8 x, char8 y);
-uchar8 __ovld __cnfn add_sat(uchar8 x, uchar8 y);
-char16 __ovld __cnfn add_sat(char16 x, char16 y);
-uchar16 __ovld __cnfn add_sat(uchar16 x, uchar16 y);
-short __ovld __cnfn add_sat(short x, short y);
-ushort __ovld __cnfn add_sat(ushort x, ushort y);
-short2 __ovld __cnfn add_sat(short2 x, short2 y);
-ushort2 __ovld __cnfn add_sat(ushort2 x, ushort2 y);
-short3 __ovld __cnfn add_sat(short3 x, short3 y);
-ushort3 __ovld __cnfn add_sat(ushort3 x, ushort3 y);
-short4 __ovld __cnfn add_sat(short4 x, short4 y);
-ushort4 __ovld __cnfn add_sat(ushort4 x, ushort4 y);
-short8 __ovld __cnfn add_sat(short8 x, short8 y);
-ushort8 __ovld __cnfn add_sat(ushort8 x, ushort8 y);
-short16 __ovld __cnfn add_sat(short16 x, short16 y);
-ushort16 __ovld __cnfn add_sat(ushort16 x, ushort16 y);
-int __ovld __cnfn add_sat(int x, int y);
-uint __ovld __cnfn add_sat(uint x, uint y);
-int2 __ovld __cnfn add_sat(int2 x, int2 y);
-uint2 __ovld __cnfn add_sat(uint2 x, uint2 y);
-int3 __ovld __cnfn add_sat(int3 x, int3 y);
-uint3 __ovld __cnfn add_sat(uint3 x, uint3 y);
-int4 __ovld __cnfn add_sat(int4 x, int4 y);
-uint4 __ovld __cnfn add_sat(uint4 x, uint4 y);
-int8 __ovld __cnfn add_sat(int8 x, int8 y);
-uint8 __ovld __cnfn add_sat(uint8 x, uint8 y);
-int16 __ovld __cnfn add_sat(int16 x, int16 y);
-uint16 __ovld __cnfn add_sat(uint16 x, uint16 y);
-long __ovld __cnfn add_sat(long x, long y);
-ulong __ovld __cnfn add_sat(ulong x, ulong y);
-long2 __ovld __cnfn add_sat(long2 x, long2 y);
-ulong2 __ovld __cnfn add_sat(ulong2 x, ulong2 y);
-long3 __ovld __cnfn add_sat(long3 x, long3 y);
-ulong3 __ovld __cnfn add_sat(ulong3 x, ulong3 y);
-long4 __ovld __cnfn add_sat(long4 x, long4 y);
-ulong4 __ovld __cnfn add_sat(ulong4 x, ulong4 y);
-long8 __ovld __cnfn add_sat(long8 x, long8 y);
-ulong8 __ovld __cnfn add_sat(ulong8 x, ulong8 y);
-long16 __ovld __cnfn add_sat(long16 x, long16 y);
-ulong16 __ovld __cnfn add_sat(ulong16 x, ulong16 y);
-
-/**
- * Returns (x + y) >> 1. The intermediate sum does
- * not modulo overflow.
- */
-char __ovld __cnfn hadd(char x, char y);
-uchar __ovld __cnfn hadd(uchar x, uchar y);
-char2 __ovld __cnfn hadd(char2 x, char2 y);
-uchar2 __ovld __cnfn hadd(uchar2 x, uchar2 y);
-char3 __ovld __cnfn hadd(char3 x, char3 y);
-uchar3 __ovld __cnfn hadd(uchar3 x, uchar3 y);
-char4 __ovld __cnfn hadd(char4 x, char4 y);
-uchar4 __ovld __cnfn hadd(uchar4 x, uchar4 y);
-char8 __ovld __cnfn hadd(char8 x, char8 y);
-uchar8 __ovld __cnfn hadd(uchar8 x, uchar8 y);
-char16 __ovld __cnfn hadd(char16 x, char16 y);
-uchar16 __ovld __cnfn hadd(uchar16 x, uchar16 y);
-short __ovld __cnfn hadd(short x, short y);
-ushort __ovld __cnfn hadd(ushort x, ushort y);
-short2 __ovld __cnfn hadd(short2 x, short2 y);
-ushort2 __ovld __cnfn hadd(ushort2 x, ushort2 y);
-short3 __ovld __cnfn hadd(short3 x, short3 y);
-ushort3 __ovld __cnfn hadd(ushort3 x, ushort3 y);
-short4 __ovld __cnfn hadd(short4 x, short4 y);
-ushort4 __ovld __cnfn hadd(ushort4 x, ushort4 y);
-short8 __ovld __cnfn hadd(short8 x, short8 y);
-ushort8 __ovld __cnfn hadd(ushort8 x, ushort8 y);
-short16 __ovld __cnfn hadd(short16 x, short16 y);
-ushort16 __ovld __cnfn hadd(ushort16 x, ushort16 y);
-int __ovld __cnfn hadd(int x, int y);
-uint __ovld __cnfn hadd(uint x, uint y);
-int2 __ovld __cnfn hadd(int2 x, int2 y);
-uint2 __ovld __cnfn hadd(uint2 x, uint2 y);
-int3 __ovld __cnfn hadd(int3 x, int3 y);
-uint3 __ovld __cnfn hadd(uint3 x, uint3 y);
-int4 __ovld __cnfn hadd(int4 x, int4 y);
-uint4 __ovld __cnfn hadd(uint4 x, uint4 y);
-int8 __ovld __cnfn hadd(int8 x, int8 y);
-uint8 __ovld __cnfn hadd(uint8 x, uint8 y);
-int16 __ovld __cnfn hadd(int16 x, int16 y);
-uint16 __ovld __cnfn hadd(uint16 x, uint16 y);
-long __ovld __cnfn hadd(long x, long y);
-ulong __ovld __cnfn hadd(ulong x, ulong y);
-long2 __ovld __cnfn hadd(long2 x, long2 y);
-ulong2 __ovld __cnfn hadd(ulong2 x, ulong2 y);
-long3 __ovld __cnfn hadd(long3 x, long3 y);
-ulong3 __ovld __cnfn hadd(ulong3 x, ulong3 y);
-long4 __ovld __cnfn hadd(long4 x, long4 y);
-ulong4 __ovld __cnfn hadd(ulong4 x, ulong4 y);
-long8 __ovld __cnfn hadd(long8 x, long8 y);
-ulong8 __ovld __cnfn hadd(ulong8 x, ulong8 y);
-long16 __ovld __cnfn hadd(long16 x, long16 y);
-ulong16 __ovld __cnfn hadd(ulong16 x, ulong16 y);
-
-/**
- * Returns (x + y + 1) >> 1. The intermediate sum
- * does not modulo overflow.
- */
-char __ovld __cnfn rhadd(char x, char y);
-uchar __ovld __cnfn rhadd(uchar x, uchar y);
-char2 __ovld __cnfn rhadd(char2 x, char2 y);
-uchar2 __ovld __cnfn rhadd(uchar2 x, uchar2 y);
-char3 __ovld __cnfn rhadd(char3 x, char3 y);
-uchar3 __ovld __cnfn rhadd(uchar3 x, uchar3 y);
-char4 __ovld __cnfn rhadd(char4 x, char4 y);
-uchar4 __ovld __cnfn rhadd(uchar4 x, uchar4 y);
-char8 __ovld __cnfn rhadd(char8 x, char8 y);
-uchar8 __ovld __cnfn rhadd(uchar8 x, uchar8 y);
-char16 __ovld __cnfn rhadd(char16 x, char16 y);
-uchar16 __ovld __cnfn rhadd(uchar16 x, uchar16 y);
-short __ovld __cnfn rhadd(short x, short y);
-ushort __ovld __cnfn rhadd(ushort x, ushort y);
-short2 __ovld __cnfn rhadd(short2 x, short2 y);
-ushort2 __ovld __cnfn rhadd(ushort2 x, ushort2 y);
-short3 __ovld __cnfn rhadd(short3 x, short3 y);
-ushort3 __ovld __cnfn rhadd(ushort3 x, ushort3 y);
-short4 __ovld __cnfn rhadd(short4 x, short4 y);
-ushort4 __ovld __cnfn rhadd(ushort4 x, ushort4 y);
-short8 __ovld __cnfn rhadd(short8 x, short8 y);
-ushort8 __ovld __cnfn rhadd(ushort8 x, ushort8 y);
-short16 __ovld __cnfn rhadd(short16 x, short16 y);
-ushort16 __ovld __cnfn rhadd(ushort16 x, ushort16 y);
-int __ovld __cnfn rhadd(int x, int y);
-uint __ovld __cnfn rhadd(uint x, uint y);
-int2 __ovld __cnfn rhadd(int2 x, int2 y);
-uint2 __ovld __cnfn rhadd(uint2 x, uint2 y);
-int3 __ovld __cnfn rhadd(int3 x, int3 y);
-uint3 __ovld __cnfn rhadd(uint3 x, uint3 y);
-int4 __ovld __cnfn rhadd(int4 x, int4 y);
-uint4 __ovld __cnfn rhadd(uint4 x, uint4 y);
-int8 __ovld __cnfn rhadd(int8 x, int8 y);
-uint8 __ovld __cnfn rhadd(uint8 x, uint8 y);
-int16 __ovld __cnfn rhadd(int16 x, int16 y);
-uint16 __ovld __cnfn rhadd(uint16 x, uint16 y);
-long __ovld __cnfn rhadd(long x, long y);
-ulong __ovld __cnfn rhadd(ulong x, ulong y);
-long2 __ovld __cnfn rhadd(long2 x, long2 y);
-ulong2 __ovld __cnfn rhadd(ulong2 x, ulong2 y);
-long3 __ovld __cnfn rhadd(long3 x, long3 y);
-ulong3 __ovld __cnfn rhadd(ulong3 x, ulong3 y);
-long4 __ovld __cnfn rhadd(long4 x, long4 y);
-ulong4 __ovld __cnfn rhadd(ulong4 x, ulong4 y);
-long8 __ovld __cnfn rhadd(long8 x, long8 y);
-ulong8 __ovld __cnfn rhadd(ulong8 x, ulong8 y);
-long16 __ovld __cnfn rhadd(long16 x, long16 y);
-ulong16 __ovld __cnfn rhadd(ulong16 x, ulong16 y);
-
-/**
- * Returns min(max(x, minval), maxval).
- * Results are undefined if minval > maxval.
- */
-char __ovld __cnfn clamp(char x, char minval, char maxval);
-uchar __ovld __cnfn clamp(uchar x, uchar minval, uchar maxval);
-char2 __ovld __cnfn clamp(char2 x, char2 minval, char2 maxval);
-uchar2 __ovld __cnfn clamp(uchar2 x, uchar2 minval, uchar2 maxval);
-char3 __ovld __cnfn clamp(char3 x, char3 minval, char3 maxval);
-uchar3 __ovld __cnfn clamp(uchar3 x, uchar3 minval, uchar3 maxval);
-char4 __ovld __cnfn clamp(char4 x, char4 minval, char4 maxval);
-uchar4 __ovld __cnfn clamp(uchar4 x, uchar4 minval, uchar4 maxval);
-char8 __ovld __cnfn clamp(char8 x, char8 minval, char8 maxval);
-uchar8 __ovld __cnfn clamp(uchar8 x, uchar8 minval, uchar8 maxval);
-char16 __ovld __cnfn clamp(char16 x, char16 minval, char16 maxval);
-uchar16 __ovld __cnfn clamp(uchar16 x, uchar16 minval, uchar16 maxval);
-short __ovld __cnfn clamp(short x, short minval, short maxval);
-ushort __ovld __cnfn clamp(ushort x, ushort minval, ushort maxval);
-short2 __ovld __cnfn clamp(short2 x, short2 minval, short2 maxval);
-ushort2 __ovld __cnfn clamp(ushort2 x, ushort2 minval, ushort2 maxval);
-short3 __ovld __cnfn clamp(short3 x, short3 minval, short3 maxval);
-ushort3 __ovld __cnfn clamp(ushort3 x, ushort3 minval, ushort3 maxval);
-short4 __ovld __cnfn clamp(short4 x, short4 minval, short4 maxval);
-ushort4 __ovld __cnfn clamp(ushort4 x, ushort4 minval, ushort4 maxval);
-short8 __ovld __cnfn clamp(short8 x, short8 minval, short8 maxval);
-ushort8 __ovld __cnfn clamp(ushort8 x, ushort8 minval, ushort8 maxval);
-short16 __ovld __cnfn clamp(short16 x, short16 minval, short16 maxval);
-ushort16 __ovld __cnfn clamp(ushort16 x, ushort16 minval, ushort16 maxval);
-int __ovld __cnfn clamp(int x, int minval, int maxval);
-uint __ovld __cnfn clamp(uint x, uint minval, uint maxval);
-int2 __ovld __cnfn clamp(int2 x, int2 minval, int2 maxval);
-uint2 __ovld __cnfn clamp(uint2 x, uint2 minval, uint2 maxval);
-int3 __ovld __cnfn clamp(int3 x, int3 minval, int3 maxval);
-uint3 __ovld __cnfn clamp(uint3 x, uint3 minval, uint3 maxval);
-int4 __ovld __cnfn clamp(int4 x, int4 minval, int4 maxval);
-uint4 __ovld __cnfn clamp(uint4 x, uint4 minval, uint4 maxval);
-int8 __ovld __cnfn clamp(int8 x, int8 minval, int8 maxval);
-uint8 __ovld __cnfn clamp(uint8 x, uint8 minval, uint8 maxval);
-int16 __ovld __cnfn clamp(int16 x, int16 minval, int16 maxval);
-uint16 __ovld __cnfn clamp(uint16 x, uint16 minval, uint16 maxval);
-long __ovld __cnfn clamp(long x, long minval, long maxval);
-ulong __ovld __cnfn clamp(ulong x, ulong minval, ulong maxval);
-long2 __ovld __cnfn clamp(long2 x, long2 minval, long2 maxval);
-ulong2 __ovld __cnfn clamp(ulong2 x, ulong2 minval, ulong2 maxval);
-long3 __ovld __cnfn clamp(long3 x, long3 minval, long3 maxval);
-ulong3 __ovld __cnfn clamp(ulong3 x, ulong3 minval, ulong3 maxval);
-long4 __ovld __cnfn clamp(long4 x, long4 minval, long4 maxval);
-ulong4 __ovld __cnfn clamp(ulong4 x, ulong4 minval, ulong4 maxval);
-long8 __ovld __cnfn clamp(long8 x, long8 minval, long8 maxval);
-ulong8 __ovld __cnfn clamp(ulong8 x, ulong8 minval, ulong8 maxval);
-long16 __ovld __cnfn clamp(long16 x, long16 minval, long16 maxval);
-ulong16 __ovld __cnfn clamp(ulong16 x, ulong16 minval, ulong16 maxval);
-char2 __ovld __cnfn clamp(char2 x, char minval, char maxval);
-uchar2 __ovld __cnfn clamp(uchar2 x, uchar minval, uchar maxval);
-char3 __ovld __cnfn clamp(char3 x, char minval, char maxval);
-uchar3 __ovld __cnfn clamp(uchar3 x, uchar minval, uchar maxval);
-char4 __ovld __cnfn clamp(char4 x, char minval, char maxval);
-uchar4 __ovld __cnfn clamp(uchar4 x, uchar minval, uchar maxval);
-char8 __ovld __cnfn clamp(char8 x, char minval, char maxval);
-uchar8 __ovld __cnfn clamp(uchar8 x, uchar minval, uchar maxval);
-char16 __ovld __cnfn clamp(char16 x, char minval, char maxval);
-uchar16 __ovld __cnfn clamp(uchar16 x, uchar minval, uchar maxval);
-short2 __ovld __cnfn clamp(short2 x, short minval, short maxval);
-ushort2 __ovld __cnfn clamp(ushort2 x, ushort minval, ushort maxval);
-short3 __ovld __cnfn clamp(short3 x, short minval, short maxval);
-ushort3 __ovld __cnfn clamp(ushort3 x, ushort minval, ushort maxval);
-short4 __ovld __cnfn clamp(short4 x, short minval, short maxval);
-ushort4 __ovld __cnfn clamp(ushort4 x, ushort minval, ushort maxval);
-short8 __ovld __cnfn clamp(short8 x, short minval, short maxval);
-ushort8 __ovld __cnfn clamp(ushort8 x, ushort minval, ushort maxval);
-short16 __ovld __cnfn clamp(short16 x, short minval, short maxval);
-ushort16 __ovld __cnfn clamp(ushort16 x, ushort minval, ushort maxval);
-int2 __ovld __cnfn clamp(int2 x, int minval, int maxval);
-uint2 __ovld __cnfn clamp(uint2 x, uint minval, uint maxval);
-int3 __ovld __cnfn clamp(int3 x, int minval, int maxval);
-uint3 __ovld __cnfn clamp(uint3 x, uint minval, uint maxval);
-int4 __ovld __cnfn clamp(int4 x, int minval, int maxval);
-uint4 __ovld __cnfn clamp(uint4 x, uint minval, uint maxval);
-int8 __ovld __cnfn clamp(int8 x, int minval, int maxval);
-uint8 __ovld __cnfn clamp(uint8 x, uint minval, uint maxval);
-int16 __ovld __cnfn clamp(int16 x, int minval, int maxval);
-uint16 __ovld __cnfn clamp(uint16 x, uint minval, uint maxval);
-long2 __ovld __cnfn clamp(long2 x, long minval, long maxval);
-ulong2 __ovld __cnfn clamp(ulong2 x, ulong minval, ulong maxval);
-long3 __ovld __cnfn clamp(long3 x, long minval, long maxval);
-ulong3 __ovld __cnfn clamp(ulong3 x, ulong minval, ulong maxval);
-long4 __ovld __cnfn clamp(long4 x, long minval, long maxval);
-ulong4 __ovld __cnfn clamp(ulong4 x, ulong minval, ulong maxval);
-long8 __ovld __cnfn clamp(long8 x, long minval, long maxval);
-ulong8 __ovld __cnfn clamp(ulong8 x, ulong minval, ulong maxval);
-long16 __ovld __cnfn clamp(long16 x, long minval, long maxval);
-ulong16 __ovld __cnfn clamp(ulong16 x, ulong minval, ulong maxval);
-
-/**
- * Returns the number of leading 0-bits in x, starting
- * at the most significant bit position.
- */
-char __ovld __cnfn clz(char x);
-uchar __ovld __cnfn clz(uchar x);
-char2 __ovld __cnfn clz(char2 x);
-uchar2 __ovld __cnfn clz(uchar2 x);
-char3 __ovld __cnfn clz(char3 x);
-uchar3 __ovld __cnfn clz(uchar3 x);
-char4 __ovld __cnfn clz(char4 x);
-uchar4 __ovld __cnfn clz(uchar4 x);
-char8 __ovld __cnfn clz(char8 x);
-uchar8 __ovld __cnfn clz(uchar8 x);
-char16 __ovld __cnfn clz(char16 x);
-uchar16 __ovld __cnfn clz(uchar16 x);
-short __ovld __cnfn clz(short x);
-ushort __ovld __cnfn clz(ushort x);
-short2 __ovld __cnfn clz(short2 x);
-ushort2 __ovld __cnfn clz(ushort2 x);
-short3 __ovld __cnfn clz(short3 x);
-ushort3 __ovld __cnfn clz(ushort3 x);
-short4 __ovld __cnfn clz(short4 x);
-ushort4 __ovld __cnfn clz(ushort4 x);
-short8 __ovld __cnfn clz(short8 x);
-ushort8 __ovld __cnfn clz(ushort8 x);
-short16 __ovld __cnfn clz(short16 x);
-ushort16 __ovld __cnfn clz(ushort16 x);
-int __ovld __cnfn clz(int x);
-uint __ovld __cnfn clz(uint x);
-int2 __ovld __cnfn clz(int2 x);
-uint2 __ovld __cnfn clz(uint2 x);
-int3 __ovld __cnfn clz(int3 x);
-uint3 __ovld __cnfn clz(uint3 x);
-int4 __ovld __cnfn clz(int4 x);
-uint4 __ovld __cnfn clz(uint4 x);
-int8 __ovld __cnfn clz(int8 x);
-uint8 __ovld __cnfn clz(uint8 x);
-int16 __ovld __cnfn clz(int16 x);
-uint16 __ovld __cnfn clz(uint16 x);
-long __ovld __cnfn clz(long x);
-ulong __ovld __cnfn clz(ulong x);
-long2 __ovld __cnfn clz(long2 x);
-ulong2 __ovld __cnfn clz(ulong2 x);
-long3 __ovld __cnfn clz(long3 x);
-ulong3 __ovld __cnfn clz(ulong3 x);
-long4 __ovld __cnfn clz(long4 x);
-ulong4 __ovld __cnfn clz(ulong4 x);
-long8 __ovld __cnfn clz(long8 x);
-ulong8 __ovld __cnfn clz(ulong8 x);
-long16 __ovld __cnfn clz(long16 x);
-ulong16 __ovld __cnfn clz(ulong16 x);
-
-/**
- * Returns the count of trailing 0-bits in x. If x is 0,
- * returns the size in bits of the type of x or
- * component type of x, if x is a vector.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-char __ovld ctz(char x);
-uchar __ovld ctz(uchar x);
-char2 __ovld ctz(char2 x);
-uchar2 __ovld ctz(uchar2 x);
-char3 __ovld ctz(char3 x);
-uchar3 __ovld ctz(uchar3 x);
-char4 __ovld ctz(char4 x);
-uchar4 __ovld ctz(uchar4 x);
-char8 __ovld ctz(char8 x);
-uchar8 __ovld ctz(uchar8 x);
-char16 __ovld ctz(char16 x);
-uchar16 __ovld ctz(uchar16 x);
-short __ovld ctz(short x);
-ushort __ovld ctz(ushort x);
-short2 __ovld ctz(short2 x);
-ushort2 __ovld ctz(ushort2 x);
-short3 __ovld ctz(short3 x);
-ushort3 __ovld ctz(ushort3 x);
-short4 __ovld ctz(short4 x);
-ushort4 __ovld ctz(ushort4 x);
-short8 __ovld ctz(short8 x);
-ushort8 __ovld ctz(ushort8 x);
-short16 __ovld ctz(short16 x);
-ushort16 __ovld ctz(ushort16 x);
-int __ovld ctz(int x);
-uint __ovld ctz(uint x);
-int2 __ovld ctz(int2 x);
-uint2 __ovld ctz(uint2 x);
-int3 __ovld ctz(int3 x);
-uint3 __ovld ctz(uint3 x);
-int4 __ovld ctz(int4 x);
-uint4 __ovld ctz(uint4 x);
-int8 __ovld ctz(int8 x);
-uint8 __ovld ctz(uint8 x);
-int16 __ovld ctz(int16 x);
-uint16 __ovld ctz(uint16 x);
-long __ovld ctz(long x);
-ulong __ovld ctz(ulong x);
-long2 __ovld ctz(long2 x);
-ulong2 __ovld ctz(ulong2 x);
-long3 __ovld ctz(long3 x);
-ulong3 __ovld ctz(ulong3 x);
-long4 __ovld ctz(long4 x);
-ulong4 __ovld ctz(ulong4 x);
-long8 __ovld ctz(long8 x);
-ulong8 __ovld ctz(ulong8 x);
-long16 __ovld ctz(long16 x);
-ulong16 __ovld ctz(ulong16 x);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Returns mul_hi(a, b) + c.
- */
-char __ovld __cnfn mad_hi(char a, char b, char c);
-uchar __ovld __cnfn mad_hi(uchar a, uchar b, uchar c);
-char2 __ovld __cnfn mad_hi(char2 a, char2 b, char2 c);
-uchar2 __ovld __cnfn mad_hi(uchar2 a, uchar2 b, uchar2 c);
-char3 __ovld __cnfn mad_hi(char3 a, char3 b, char3 c);
-uchar3 __ovld __cnfn mad_hi(uchar3 a, uchar3 b, uchar3 c);
-char4 __ovld __cnfn mad_hi(char4 a, char4 b, char4 c);
-uchar4 __ovld __cnfn mad_hi(uchar4 a, uchar4 b, uchar4 c);
-char8 __ovld __cnfn mad_hi(char8 a, char8 b, char8 c);
-uchar8 __ovld __cnfn mad_hi(uchar8 a, uchar8 b, uchar8 c);
-char16 __ovld __cnfn mad_hi(char16 a, char16 b, char16 c);
-uchar16 __ovld __cnfn mad_hi(uchar16 a, uchar16 b, uchar16 c);
-short __ovld __cnfn mad_hi(short a, short b, short c);
-ushort __ovld __cnfn mad_hi(ushort a, ushort b, ushort c);
-short2 __ovld __cnfn mad_hi(short2 a, short2 b, short2 c);
-ushort2 __ovld __cnfn mad_hi(ushort2 a, ushort2 b, ushort2 c);
-short3 __ovld __cnfn mad_hi(short3 a, short3 b, short3 c);
-ushort3 __ovld __cnfn mad_hi(ushort3 a, ushort3 b, ushort3 c);
-short4 __ovld __cnfn mad_hi(short4 a, short4 b, short4 c);
-ushort4 __ovld __cnfn mad_hi(ushort4 a, ushort4 b, ushort4 c);
-short8 __ovld __cnfn mad_hi(short8 a, short8 b, short8 c);
-ushort8 __ovld __cnfn mad_hi(ushort8 a, ushort8 b, ushort8 c);
-short16 __ovld __cnfn mad_hi(short16 a, short16 b, short16 c);
-ushort16 __ovld __cnfn mad_hi(ushort16 a, ushort16 b, ushort16 c);
-int __ovld __cnfn mad_hi(int a, int b, int c);
-uint __ovld __cnfn mad_hi(uint a, uint b, uint c);
-int2 __ovld __cnfn mad_hi(int2 a, int2 b, int2 c);
-uint2 __ovld __cnfn mad_hi(uint2 a, uint2 b, uint2 c);
-int3 __ovld __cnfn mad_hi(int3 a, int3 b, int3 c);
-uint3 __ovld __cnfn mad_hi(uint3 a, uint3 b, uint3 c);
-int4 __ovld __cnfn mad_hi(int4 a, int4 b, int4 c);
-uint4 __ovld __cnfn mad_hi(uint4 a, uint4 b, uint4 c);
-int8 __ovld __cnfn mad_hi(int8 a, int8 b, int8 c);
-uint8 __ovld __cnfn mad_hi(uint8 a, uint8 b, uint8 c);
-int16 __ovld __cnfn mad_hi(int16 a, int16 b, int16 c);
-uint16 __ovld __cnfn mad_hi(uint16 a, uint16 b, uint16 c);
-long __ovld __cnfn mad_hi(long a, long b, long c);
-ulong __ovld __cnfn mad_hi(ulong a, ulong b, ulong c);
-long2 __ovld __cnfn mad_hi(long2 a, long2 b, long2 c);
-ulong2 __ovld __cnfn mad_hi(ulong2 a, ulong2 b, ulong2 c);
-long3 __ovld __cnfn mad_hi(long3 a, long3 b, long3 c);
-ulong3 __ovld __cnfn mad_hi(ulong3 a, ulong3 b, ulong3 c);
-long4 __ovld __cnfn mad_hi(long4 a, long4 b, long4 c);
-ulong4 __ovld __cnfn mad_hi(ulong4 a, ulong4 b, ulong4 c);
-long8 __ovld __cnfn mad_hi(long8 a, long8 b, long8 c);
-ulong8 __ovld __cnfn mad_hi(ulong8 a, ulong8 b, ulong8 c);
-long16 __ovld __cnfn mad_hi(long16 a, long16 b, long16 c);
-ulong16 __ovld __cnfn mad_hi(ulong16 a, ulong16 b, ulong16 c);
-
-/**
- * Returns a * b + c and saturates the result.
- */
-char __ovld __cnfn mad_sat(char a, char b, char c);
-uchar __ovld __cnfn mad_sat(uchar a, uchar b, uchar c);
-char2 __ovld __cnfn mad_sat(char2 a, char2 b, char2 c);
-uchar2 __ovld __cnfn mad_sat(uchar2 a, uchar2 b, uchar2 c);
-char3 __ovld __cnfn mad_sat(char3 a, char3 b, char3 c);
-uchar3 __ovld __cnfn mad_sat(uchar3 a, uchar3 b, uchar3 c);
-char4 __ovld __cnfn mad_sat(char4 a, char4 b, char4 c);
-uchar4 __ovld __cnfn mad_sat(uchar4 a, uchar4 b, uchar4 c);
-char8 __ovld __cnfn mad_sat(char8 a, char8 b, char8 c);
-uchar8 __ovld __cnfn mad_sat(uchar8 a, uchar8 b, uchar8 c);
-char16 __ovld __cnfn mad_sat(char16 a, char16 b, char16 c);
-uchar16 __ovld __cnfn mad_sat(uchar16 a, uchar16 b, uchar16 c);
-short __ovld __cnfn mad_sat(short a, short b, short c);
-ushort __ovld __cnfn mad_sat(ushort a, ushort b, ushort c);
-short2 __ovld __cnfn mad_sat(short2 a, short2 b, short2 c);
-ushort2 __ovld __cnfn mad_sat(ushort2 a, ushort2 b, ushort2 c);
-short3 __ovld __cnfn mad_sat(short3 a, short3 b, short3 c);
-ushort3 __ovld __cnfn mad_sat(ushort3 a, ushort3 b, ushort3 c);
-short4 __ovld __cnfn mad_sat(short4 a, short4 b, short4 c);
-ushort4 __ovld __cnfn mad_sat(ushort4 a, ushort4 b, ushort4 c);
-short8 __ovld __cnfn mad_sat(short8 a, short8 b, short8 c);
-ushort8 __ovld __cnfn mad_sat(ushort8 a, ushort8 b, ushort8 c);
-short16 __ovld __cnfn mad_sat(short16 a, short16 b, short16 c);
-ushort16 __ovld __cnfn mad_sat(ushort16 a, ushort16 b, ushort16 c);
-int __ovld __cnfn mad_sat(int a, int b, int c);
-uint __ovld __cnfn mad_sat(uint a, uint b, uint c);
-int2 __ovld __cnfn mad_sat(int2 a, int2 b, int2 c);
-uint2 __ovld __cnfn mad_sat(uint2 a, uint2 b, uint2 c);
-int3 __ovld __cnfn mad_sat(int3 a, int3 b, int3 c);
-uint3 __ovld __cnfn mad_sat(uint3 a, uint3 b, uint3 c);
-int4 __ovld __cnfn mad_sat(int4 a, int4 b, int4 c);
-uint4 __ovld __cnfn mad_sat(uint4 a, uint4 b, uint4 c);
-int8 __ovld __cnfn mad_sat(int8 a, int8 b, int8 c);
-uint8 __ovld __cnfn mad_sat(uint8 a, uint8 b, uint8 c);
-int16 __ovld __cnfn mad_sat(int16 a, int16 b, int16 c);
-uint16 __ovld __cnfn mad_sat(uint16 a, uint16 b, uint16 c);
-long __ovld __cnfn mad_sat(long a, long b, long c);
-ulong __ovld __cnfn mad_sat(ulong a, ulong b, ulong c);
-long2 __ovld __cnfn mad_sat(long2 a, long2 b, long2 c);
-ulong2 __ovld __cnfn mad_sat(ulong2 a, ulong2 b, ulong2 c);
-long3 __ovld __cnfn mad_sat(long3 a, long3 b, long3 c);
-ulong3 __ovld __cnfn mad_sat(ulong3 a, ulong3 b, ulong3 c);
-long4 __ovld __cnfn mad_sat(long4 a, long4 b, long4 c);
-ulong4 __ovld __cnfn mad_sat(ulong4 a, ulong4 b, ulong4 c);
-long8 __ovld __cnfn mad_sat(long8 a, long8 b, long8 c);
-ulong8 __ovld __cnfn mad_sat(ulong8 a, ulong8 b, ulong8 c);
-long16 __ovld __cnfn mad_sat(long16 a, long16 b, long16 c);
-ulong16 __ovld __cnfn mad_sat(ulong16 a, ulong16 b, ulong16 c);
-
-/**
- * Returns y if x < y, otherwise it returns x.
- */
-char __ovld __cnfn max(char x, char y);
-uchar __ovld __cnfn max(uchar x, uchar y);
-char2 __ovld __cnfn max(char2 x, char2 y);
-uchar2 __ovld __cnfn max(uchar2 x, uchar2 y);
-char3 __ovld __cnfn max(char3 x, char3 y);
-uchar3 __ovld __cnfn max(uchar3 x, uchar3 y);
-char4 __ovld __cnfn max(char4 x, char4 y);
-uchar4 __ovld __cnfn max(uchar4 x, uchar4 y);
-char8 __ovld __cnfn max(char8 x, char8 y);
-uchar8 __ovld __cnfn max(uchar8 x, uchar8 y);
-char16 __ovld __cnfn max(char16 x, char16 y);
-uchar16 __ovld __cnfn max(uchar16 x, uchar16 y);
-short __ovld __cnfn max(short x, short y);
-ushort __ovld __cnfn max(ushort x, ushort y);
-short2 __ovld __cnfn max(short2 x, short2 y);
-ushort2 __ovld __cnfn max(ushort2 x, ushort2 y);
-short3 __ovld __cnfn max(short3 x, short3 y);
-ushort3 __ovld __cnfn max(ushort3 x, ushort3 y);
-short4 __ovld __cnfn max(short4 x, short4 y);
-ushort4 __ovld __cnfn max(ushort4 x, ushort4 y);
-short8 __ovld __cnfn max(short8 x, short8 y);
-ushort8 __ovld __cnfn max(ushort8 x, ushort8 y);
-short16 __ovld __cnfn max(short16 x, short16 y);
-ushort16 __ovld __cnfn max(ushort16 x, ushort16 y);
-int __ovld __cnfn max(int x, int y);
-uint __ovld __cnfn max(uint x, uint y);
-int2 __ovld __cnfn max(int2 x, int2 y);
-uint2 __ovld __cnfn max(uint2 x, uint2 y);
-int3 __ovld __cnfn max(int3 x, int3 y);
-uint3 __ovld __cnfn max(uint3 x, uint3 y);
-int4 __ovld __cnfn max(int4 x, int4 y);
-uint4 __ovld __cnfn max(uint4 x, uint4 y);
-int8 __ovld __cnfn max(int8 x, int8 y);
-uint8 __ovld __cnfn max(uint8 x, uint8 y);
-int16 __ovld __cnfn max(int16 x, int16 y);
-uint16 __ovld __cnfn max(uint16 x, uint16 y);
-long __ovld __cnfn max(long x, long y);
-ulong __ovld __cnfn max(ulong x, ulong y);
-long2 __ovld __cnfn max(long2 x, long2 y);
-ulong2 __ovld __cnfn max(ulong2 x, ulong2 y);
-long3 __ovld __cnfn max(long3 x, long3 y);
-ulong3 __ovld __cnfn max(ulong3 x, ulong3 y);
-long4 __ovld __cnfn max(long4 x, long4 y);
-ulong4 __ovld __cnfn max(ulong4 x, ulong4 y);
-long8 __ovld __cnfn max(long8 x, long8 y);
-ulong8 __ovld __cnfn max(ulong8 x, ulong8 y);
-long16 __ovld __cnfn max(long16 x, long16 y);
-ulong16 __ovld __cnfn max(ulong16 x, ulong16 y);
-char2 __ovld __cnfn max(char2 x, char y);
-uchar2 __ovld __cnfn max(uchar2 x, uchar y);
-char3 __ovld __cnfn max(char3 x, char y);
-uchar3 __ovld __cnfn max(uchar3 x, uchar y);
-char4 __ovld __cnfn max(char4 x, char y);
-uchar4 __ovld __cnfn max(uchar4 x, uchar y);
-char8 __ovld __cnfn max(char8 x, char y);
-uchar8 __ovld __cnfn max(uchar8 x, uchar y);
-char16 __ovld __cnfn max(char16 x, char y);
-uchar16 __ovld __cnfn max(uchar16 x, uchar y);
-short2 __ovld __cnfn max(short2 x, short y);
-ushort2 __ovld __cnfn max(ushort2 x, ushort y);
-short3 __ovld __cnfn max(short3 x, short y);
-ushort3 __ovld __cnfn max(ushort3 x, ushort y);
-short4 __ovld __cnfn max(short4 x, short y);
-ushort4 __ovld __cnfn max(ushort4 x, ushort y);
-short8 __ovld __cnfn max(short8 x, short y);
-ushort8 __ovld __cnfn max(ushort8 x, ushort y);
-short16 __ovld __cnfn max(short16 x, short y);
-ushort16 __ovld __cnfn max(ushort16 x, ushort y);
-int2 __ovld __cnfn max(int2 x, int y);
-uint2 __ovld __cnfn max(uint2 x, uint y);
-int3 __ovld __cnfn max(int3 x, int y);
-uint3 __ovld __cnfn max(uint3 x, uint y);
-int4 __ovld __cnfn max(int4 x, int y);
-uint4 __ovld __cnfn max(uint4 x, uint y);
-int8 __ovld __cnfn max(int8 x, int y);
-uint8 __ovld __cnfn max(uint8 x, uint y);
-int16 __ovld __cnfn max(int16 x, int y);
-uint16 __ovld __cnfn max(uint16 x, uint y);
-long2 __ovld __cnfn max(long2 x, long y);
-ulong2 __ovld __cnfn max(ulong2 x, ulong y);
-long3 __ovld __cnfn max(long3 x, long y);
-ulong3 __ovld __cnfn max(ulong3 x, ulong y);
-long4 __ovld __cnfn max(long4 x, long y);
-ulong4 __ovld __cnfn max(ulong4 x, ulong y);
-long8 __ovld __cnfn max(long8 x, long y);
-ulong8 __ovld __cnfn max(ulong8 x, ulong y);
-long16 __ovld __cnfn max(long16 x, long y);
-ulong16 __ovld __cnfn max(ulong16 x, ulong y);
-
-/**
- * Returns y if y < x, otherwise it returns x.
- */
-char __ovld __cnfn min(char x, char y);
-uchar __ovld __cnfn min(uchar x, uchar y);
-char2 __ovld __cnfn min(char2 x, char2 y);
-uchar2 __ovld __cnfn min(uchar2 x, uchar2 y);
-char3 __ovld __cnfn min(char3 x, char3 y);
-uchar3 __ovld __cnfn min(uchar3 x, uchar3 y);
-char4 __ovld __cnfn min(char4 x, char4 y);
-uchar4 __ovld __cnfn min(uchar4 x, uchar4 y);
-char8 __ovld __cnfn min(char8 x, char8 y);
-uchar8 __ovld __cnfn min(uchar8 x, uchar8 y);
-char16 __ovld __cnfn min(char16 x, char16 y);
-uchar16 __ovld __cnfn min(uchar16 x, uchar16 y);
-short __ovld __cnfn min(short x, short y);
-ushort __ovld __cnfn min(ushort x, ushort y);
-short2 __ovld __cnfn min(short2 x, short2 y);
-ushort2 __ovld __cnfn min(ushort2 x, ushort2 y);
-short3 __ovld __cnfn min(short3 x, short3 y);
-ushort3 __ovld __cnfn min(ushort3 x, ushort3 y);
-short4 __ovld __cnfn min(short4 x, short4 y);
-ushort4 __ovld __cnfn min(ushort4 x, ushort4 y);
-short8 __ovld __cnfn min(short8 x, short8 y);
-ushort8 __ovld __cnfn min(ushort8 x, ushort8 y);
-short16 __ovld __cnfn min(short16 x, short16 y);
-ushort16 __ovld __cnfn min(ushort16 x, ushort16 y);
-int __ovld __cnfn min(int x, int y);
-uint __ovld __cnfn min(uint x, uint y);
-int2 __ovld __cnfn min(int2 x, int2 y);
-uint2 __ovld __cnfn min(uint2 x, uint2 y);
-int3 __ovld __cnfn min(int3 x, int3 y);
-uint3 __ovld __cnfn min(uint3 x, uint3 y);
-int4 __ovld __cnfn min(int4 x, int4 y);
-uint4 __ovld __cnfn min(uint4 x, uint4 y);
-int8 __ovld __cnfn min(int8 x, int8 y);
-uint8 __ovld __cnfn min(uint8 x, uint8 y);
-int16 __ovld __cnfn min(int16 x, int16 y);
-uint16 __ovld __cnfn min(uint16 x, uint16 y);
-long __ovld __cnfn min(long x, long y);
-ulong __ovld __cnfn min(ulong x, ulong y);
-long2 __ovld __cnfn min(long2 x, long2 y);
-ulong2 __ovld __cnfn min(ulong2 x, ulong2 y);
-long3 __ovld __cnfn min(long3 x, long3 y);
-ulong3 __ovld __cnfn min(ulong3 x, ulong3 y);
-long4 __ovld __cnfn min(long4 x, long4 y);
-ulong4 __ovld __cnfn min(ulong4 x, ulong4 y);
-long8 __ovld __cnfn min(long8 x, long8 y);
-ulong8 __ovld __cnfn min(ulong8 x, ulong8 y);
-long16 __ovld __cnfn min(long16 x, long16 y);
-ulong16 __ovld __cnfn min(ulong16 x, ulong16 y);
-char2 __ovld __cnfn min(char2 x, char y);
-uchar2 __ovld __cnfn min(uchar2 x, uchar y);
-char3 __ovld __cnfn min(char3 x, char y);
-uchar3 __ovld __cnfn min(uchar3 x, uchar y);
-char4 __ovld __cnfn min(char4 x, char y);
-uchar4 __ovld __cnfn min(uchar4 x, uchar y);
-char8 __ovld __cnfn min(char8 x, char y);
-uchar8 __ovld __cnfn min(uchar8 x, uchar y);
-char16 __ovld __cnfn min(char16 x, char y);
-uchar16 __ovld __cnfn min(uchar16 x, uchar y);
-short2 __ovld __cnfn min(short2 x, short y);
-ushort2 __ovld __cnfn min(ushort2 x, ushort y);
-short3 __ovld __cnfn min(short3 x, short y);
-ushort3 __ovld __cnfn min(ushort3 x, ushort y);
-short4 __ovld __cnfn min(short4 x, short y);
-ushort4 __ovld __cnfn min(ushort4 x, ushort y);
-short8 __ovld __cnfn min(short8 x, short y);
-ushort8 __ovld __cnfn min(ushort8 x, ushort y);
-short16 __ovld __cnfn min(short16 x, short y);
-ushort16 __ovld __cnfn min(ushort16 x, ushort y);
-int2 __ovld __cnfn min(int2 x, int y);
-uint2 __ovld __cnfn min(uint2 x, uint y);
-int3 __ovld __cnfn min(int3 x, int y);
-uint3 __ovld __cnfn min(uint3 x, uint y);
-int4 __ovld __cnfn min(int4 x, int y);
-uint4 __ovld __cnfn min(uint4 x, uint y);
-int8 __ovld __cnfn min(int8 x, int y);
-uint8 __ovld __cnfn min(uint8 x, uint y);
-int16 __ovld __cnfn min(int16 x, int y);
-uint16 __ovld __cnfn min(uint16 x, uint y);
-long2 __ovld __cnfn min(long2 x, long y);
-ulong2 __ovld __cnfn min(ulong2 x, ulong y);
-long3 __ovld __cnfn min(long3 x, long y);
-ulong3 __ovld __cnfn min(ulong3 x, ulong y);
-long4 __ovld __cnfn min(long4 x, long y);
-ulong4 __ovld __cnfn min(ulong4 x, ulong y);
-long8 __ovld __cnfn min(long8 x, long y);
-ulong8 __ovld __cnfn min(ulong8 x, ulong y);
-long16 __ovld __cnfn min(long16 x, long y);
-ulong16 __ovld __cnfn min(ulong16 x, ulong y);
-
-/**
- * Computes x * y and returns the high half of the
- * product of x and y.
- */
-char __ovld __cnfn mul_hi(char x, char y);
-uchar __ovld __cnfn mul_hi(uchar x, uchar y);
-char2 __ovld __cnfn mul_hi(char2 x, char2 y);
-uchar2 __ovld __cnfn mul_hi(uchar2 x, uchar2 y);
-char3 __ovld __cnfn mul_hi(char3 x, char3 y);
-uchar3 __ovld __cnfn mul_hi(uchar3 x, uchar3 y);
-char4 __ovld __cnfn mul_hi(char4 x, char4 y);
-uchar4 __ovld __cnfn mul_hi(uchar4 x, uchar4 y);
-char8 __ovld __cnfn mul_hi(char8 x, char8 y);
-uchar8 __ovld __cnfn mul_hi(uchar8 x, uchar8 y);
-char16 __ovld __cnfn mul_hi(char16 x, char16 y);
-uchar16 __ovld __cnfn mul_hi(uchar16 x, uchar16 y);
-short __ovld __cnfn mul_hi(short x, short y);
-ushort __ovld __cnfn mul_hi(ushort x, ushort y);
-short2 __ovld __cnfn mul_hi(short2 x, short2 y);
-ushort2 __ovld __cnfn mul_hi(ushort2 x, ushort2 y);
-short3 __ovld __cnfn mul_hi(short3 x, short3 y);
-ushort3 __ovld __cnfn mul_hi(ushort3 x, ushort3 y);
-short4 __ovld __cnfn mul_hi(short4 x, short4 y);
-ushort4 __ovld __cnfn mul_hi(ushort4 x, ushort4 y);
-short8 __ovld __cnfn mul_hi(short8 x, short8 y);
-ushort8 __ovld __cnfn mul_hi(ushort8 x, ushort8 y);
-short16 __ovld __cnfn mul_hi(short16 x, short16 y);
-ushort16 __ovld __cnfn mul_hi(ushort16 x, ushort16 y);
-int __ovld __cnfn mul_hi(int x, int y);
-uint __ovld __cnfn mul_hi(uint x, uint y);
-int2 __ovld __cnfn mul_hi(int2 x, int2 y);
-uint2 __ovld __cnfn mul_hi(uint2 x, uint2 y);
-int3 __ovld __cnfn mul_hi(int3 x, int3 y);
-uint3 __ovld __cnfn mul_hi(uint3 x, uint3 y);
-int4 __ovld __cnfn mul_hi(int4 x, int4 y);
-uint4 __ovld __cnfn mul_hi(uint4 x, uint4 y);
-int8 __ovld __cnfn mul_hi(int8 x, int8 y);
-uint8 __ovld __cnfn mul_hi(uint8 x, uint8 y);
-int16 __ovld __cnfn mul_hi(int16 x, int16 y);
-uint16 __ovld __cnfn mul_hi(uint16 x, uint16 y);
-long __ovld __cnfn mul_hi(long x, long y);
-ulong __ovld __cnfn mul_hi(ulong x, ulong y);
-long2 __ovld __cnfn mul_hi(long2 x, long2 y);
-ulong2 __ovld __cnfn mul_hi(ulong2 x, ulong2 y);
-long3 __ovld __cnfn mul_hi(long3 x, long3 y);
-ulong3 __ovld __cnfn mul_hi(ulong3 x, ulong3 y);
-long4 __ovld __cnfn mul_hi(long4 x, long4 y);
-ulong4 __ovld __cnfn mul_hi(ulong4 x, ulong4 y);
-long8 __ovld __cnfn mul_hi(long8 x, long8 y);
-ulong8 __ovld __cnfn mul_hi(ulong8 x, ulong8 y);
-long16 __ovld __cnfn mul_hi(long16 x, long16 y);
-ulong16 __ovld __cnfn mul_hi(ulong16 x, ulong16 y);
-
-/**
- * For each element in v, the bits are shifted left by
- * the number of bits given by the corresponding
- * element in i (subject to usual shift modulo rules
- * described in section 6.3). Bits shifted off the left
- * side of the element are shifted back in from the
- * right.
- */
-char __ovld __cnfn rotate(char v, char i);
-uchar __ovld __cnfn rotate(uchar v, uchar i);
-char2 __ovld __cnfn rotate(char2 v, char2 i);
-uchar2 __ovld __cnfn rotate(uchar2 v, uchar2 i);
-char3 __ovld __cnfn rotate(char3 v, char3 i);
-uchar3 __ovld __cnfn rotate(uchar3 v, uchar3 i);
-char4 __ovld __cnfn rotate(char4 v, char4 i);
-uchar4 __ovld __cnfn rotate(uchar4 v, uchar4 i);
-char8 __ovld __cnfn rotate(char8 v, char8 i);
-uchar8 __ovld __cnfn rotate(uchar8 v, uchar8 i);
-char16 __ovld __cnfn rotate(char16 v, char16 i);
-uchar16 __ovld __cnfn rotate(uchar16 v, uchar16 i);
-short __ovld __cnfn rotate(short v, short i);
-ushort __ovld __cnfn rotate(ushort v, ushort i);
-short2 __ovld __cnfn rotate(short2 v, short2 i);
-ushort2 __ovld __cnfn rotate(ushort2 v, ushort2 i);
-short3 __ovld __cnfn rotate(short3 v, short3 i);
-ushort3 __ovld __cnfn rotate(ushort3 v, ushort3 i);
-short4 __ovld __cnfn rotate(short4 v, short4 i);
-ushort4 __ovld __cnfn rotate(ushort4 v, ushort4 i);
-short8 __ovld __cnfn rotate(short8 v, short8 i);
-ushort8 __ovld __cnfn rotate(ushort8 v, ushort8 i);
-short16 __ovld __cnfn rotate(short16 v, short16 i);
-ushort16 __ovld __cnfn rotate(ushort16 v, ushort16 i);
-int __ovld __cnfn rotate(int v, int i);
-uint __ovld __cnfn rotate(uint v, uint i);
-int2 __ovld __cnfn rotate(int2 v, int2 i);
-uint2 __ovld __cnfn rotate(uint2 v, uint2 i);
-int3 __ovld __cnfn rotate(int3 v, int3 i);
-uint3 __ovld __cnfn rotate(uint3 v, uint3 i);
-int4 __ovld __cnfn rotate(int4 v, int4 i);
-uint4 __ovld __cnfn rotate(uint4 v, uint4 i);
-int8 __ovld __cnfn rotate(int8 v, int8 i);
-uint8 __ovld __cnfn rotate(uint8 v, uint8 i);
-int16 __ovld __cnfn rotate(int16 v, int16 i);
-uint16 __ovld __cnfn rotate(uint16 v, uint16 i);
-long __ovld __cnfn rotate(long v, long i);
-ulong __ovld __cnfn rotate(ulong v, ulong i);
-long2 __ovld __cnfn rotate(long2 v, long2 i);
-ulong2 __ovld __cnfn rotate(ulong2 v, ulong2 i);
-long3 __ovld __cnfn rotate(long3 v, long3 i);
-ulong3 __ovld __cnfn rotate(ulong3 v, ulong3 i);
-long4 __ovld __cnfn rotate(long4 v, long4 i);
-ulong4 __ovld __cnfn rotate(ulong4 v, ulong4 i);
-long8 __ovld __cnfn rotate(long8 v, long8 i);
-ulong8 __ovld __cnfn rotate(ulong8 v, ulong8 i);
-long16 __ovld __cnfn rotate(long16 v, long16 i);
-ulong16 __ovld __cnfn rotate(ulong16 v, ulong16 i);
-
-/**
- * Returns x - y and saturates the result.
- */
-char __ovld __cnfn sub_sat(char x, char y);
-uchar __ovld __cnfn sub_sat(uchar x, uchar y);
-char2 __ovld __cnfn sub_sat(char2 x, char2 y);
-uchar2 __ovld __cnfn sub_sat(uchar2 x, uchar2 y);
-char3 __ovld __cnfn sub_sat(char3 x, char3 y);
-uchar3 __ovld __cnfn sub_sat(uchar3 x, uchar3 y);
-char4 __ovld __cnfn sub_sat(char4 x, char4 y);
-uchar4 __ovld __cnfn sub_sat(uchar4 x, uchar4 y);
-char8 __ovld __cnfn sub_sat(char8 x, char8 y);
-uchar8 __ovld __cnfn sub_sat(uchar8 x, uchar8 y);
-char16 __ovld __cnfn sub_sat(char16 x, char16 y);
-uchar16 __ovld __cnfn sub_sat(uchar16 x, uchar16 y);
-short __ovld __cnfn sub_sat(short x, short y);
-ushort __ovld __cnfn sub_sat(ushort x, ushort y);
-short2 __ovld __cnfn sub_sat(short2 x, short2 y);
-ushort2 __ovld __cnfn sub_sat(ushort2 x, ushort2 y);
-short3 __ovld __cnfn sub_sat(short3 x, short3 y);
-ushort3 __ovld __cnfn sub_sat(ushort3 x, ushort3 y);
-short4 __ovld __cnfn sub_sat(short4 x, short4 y);
-ushort4 __ovld __cnfn sub_sat(ushort4 x, ushort4 y);
-short8 __ovld __cnfn sub_sat(short8 x, short8 y);
-ushort8 __ovld __cnfn sub_sat(ushort8 x, ushort8 y);
-short16 __ovld __cnfn sub_sat(short16 x, short16 y);
-ushort16 __ovld __cnfn sub_sat(ushort16 x, ushort16 y);
-int __ovld __cnfn sub_sat(int x, int y);
-uint __ovld __cnfn sub_sat(uint x, uint y);
-int2 __ovld __cnfn sub_sat(int2 x, int2 y);
-uint2 __ovld __cnfn sub_sat(uint2 x, uint2 y);
-int3 __ovld __cnfn sub_sat(int3 x, int3 y);
-uint3 __ovld __cnfn sub_sat(uint3 x, uint3 y);
-int4 __ovld __cnfn sub_sat(int4 x, int4 y);
-uint4 __ovld __cnfn sub_sat(uint4 x, uint4 y);
-int8 __ovld __cnfn sub_sat(int8 x, int8 y);
-uint8 __ovld __cnfn sub_sat(uint8 x, uint8 y);
-int16 __ovld __cnfn sub_sat(int16 x, int16 y);
-uint16 __ovld __cnfn sub_sat(uint16 x, uint16 y);
-long __ovld __cnfn sub_sat(long x, long y);
-ulong __ovld __cnfn sub_sat(ulong x, ulong y);
-long2 __ovld __cnfn sub_sat(long2 x, long2 y);
-ulong2 __ovld __cnfn sub_sat(ulong2 x, ulong2 y);
-long3 __ovld __cnfn sub_sat(long3 x, long3 y);
-ulong3 __ovld __cnfn sub_sat(ulong3 x, ulong3 y);
-long4 __ovld __cnfn sub_sat(long4 x, long4 y);
-ulong4 __ovld __cnfn sub_sat(ulong4 x, ulong4 y);
-long8 __ovld __cnfn sub_sat(long8 x, long8 y);
-ulong8 __ovld __cnfn sub_sat(ulong8 x, ulong8 y);
-long16 __ovld __cnfn sub_sat(long16 x, long16 y);
-ulong16 __ovld __cnfn sub_sat(ulong16 x, ulong16 y);
-
-/**
- * result[i] = ((short)hi[i] << 8) | lo[i]
- * result[i] = ((ushort)hi[i] << 8) | lo[i]
- */
-short __ovld __cnfn upsample(char hi, uchar lo);
-ushort __ovld __cnfn upsample(uchar hi, uchar lo);
-short2 __ovld __cnfn upsample(char2 hi, uchar2 lo);
-short3 __ovld __cnfn upsample(char3 hi, uchar3 lo);
-short4 __ovld __cnfn upsample(char4 hi, uchar4 lo);
-short8 __ovld __cnfn upsample(char8 hi, uchar8 lo);
-short16 __ovld __cnfn upsample(char16 hi, uchar16 lo);
-ushort2 __ovld __cnfn upsample(uchar2 hi, uchar2 lo);
-ushort3 __ovld __cnfn upsample(uchar3 hi, uchar3 lo);
-ushort4 __ovld __cnfn upsample(uchar4 hi, uchar4 lo);
-ushort8 __ovld __cnfn upsample(uchar8 hi, uchar8 lo);
-ushort16 __ovld __cnfn upsample(uchar16 hi, uchar16 lo);
-
-/**
- * result[i] = ((int)hi[i] << 16) | lo[i]
- * result[i] = ((uint)hi[i] << 16) | lo[i]
- */
-int __ovld __cnfn upsample(short hi, ushort lo);
-uint __ovld __cnfn upsample(ushort hi, ushort lo);
-int2 __ovld __cnfn upsample(short2 hi, ushort2 lo);
-int3 __ovld __cnfn upsample(short3 hi, ushort3 lo);
-int4 __ovld __cnfn upsample(short4 hi, ushort4 lo);
-int8 __ovld __cnfn upsample(short8 hi, ushort8 lo);
-int16 __ovld __cnfn upsample(short16 hi, ushort16 lo);
-uint2 __ovld __cnfn upsample(ushort2 hi, ushort2 lo);
-uint3 __ovld __cnfn upsample(ushort3 hi, ushort3 lo);
-uint4 __ovld __cnfn upsample(ushort4 hi, ushort4 lo);
-uint8 __ovld __cnfn upsample(ushort8 hi, ushort8 lo);
-uint16 __ovld __cnfn upsample(ushort16 hi, ushort16 lo);
-/**
- * result[i] = ((long)hi[i] << 32) | lo[i]
- * result[i] = ((ulong)hi[i] << 32) | lo[i]
- */
-long __ovld __cnfn upsample(int hi, uint lo);
-ulong __ovld __cnfn upsample(uint hi, uint lo);
-long2 __ovld __cnfn upsample(int2 hi, uint2 lo);
-long3 __ovld __cnfn upsample(int3 hi, uint3 lo);
-long4 __ovld __cnfn upsample(int4 hi, uint4 lo);
-long8 __ovld __cnfn upsample(int8 hi, uint8 lo);
-long16 __ovld __cnfn upsample(int16 hi, uint16 lo);
-ulong2 __ovld __cnfn upsample(uint2 hi, uint2 lo);
-ulong3 __ovld __cnfn upsample(uint3 hi, uint3 lo);
-ulong4 __ovld __cnfn upsample(uint4 hi, uint4 lo);
-ulong8 __ovld __cnfn upsample(uint8 hi, uint8 lo);
-ulong16 __ovld __cnfn upsample(uint16 hi, uint16 lo);
-
-/*
- * popcount(x): returns the number of set bit in x
- */
-char __ovld __cnfn popcount(char x);
-uchar __ovld __cnfn popcount(uchar x);
-char2 __ovld __cnfn popcount(char2 x);
-uchar2 __ovld __cnfn popcount(uchar2 x);
-char3 __ovld __cnfn popcount(char3 x);
-uchar3 __ovld __cnfn popcount(uchar3 x);
-char4 __ovld __cnfn popcount(char4 x);
-uchar4 __ovld __cnfn popcount(uchar4 x);
-char8 __ovld __cnfn popcount(char8 x);
-uchar8 __ovld __cnfn popcount(uchar8 x);
-char16 __ovld __cnfn popcount(char16 x);
-uchar16 __ovld __cnfn popcount(uchar16 x);
-short __ovld __cnfn popcount(short x);
-ushort __ovld __cnfn popcount(ushort x);
-short2 __ovld __cnfn popcount(short2 x);
-ushort2 __ovld __cnfn popcount(ushort2 x);
-short3 __ovld __cnfn popcount(short3 x);
-ushort3 __ovld __cnfn popcount(ushort3 x);
-short4 __ovld __cnfn popcount(short4 x);
-ushort4 __ovld __cnfn popcount(ushort4 x);
-short8 __ovld __cnfn popcount(short8 x);
-ushort8 __ovld __cnfn popcount(ushort8 x);
-short16 __ovld __cnfn popcount(short16 x);
-ushort16 __ovld __cnfn popcount(ushort16 x);
-int __ovld __cnfn popcount(int x);
-uint __ovld __cnfn popcount(uint x);
-int2 __ovld __cnfn popcount(int2 x);
-uint2 __ovld __cnfn popcount(uint2 x);
-int3 __ovld __cnfn popcount(int3 x);
-uint3 __ovld __cnfn popcount(uint3 x);
-int4 __ovld __cnfn popcount(int4 x);
-uint4 __ovld __cnfn popcount(uint4 x);
-int8 __ovld __cnfn popcount(int8 x);
-uint8 __ovld __cnfn popcount(uint8 x);
-int16 __ovld __cnfn popcount(int16 x);
-uint16 __ovld __cnfn popcount(uint16 x);
-long __ovld __cnfn popcount(long x);
-ulong __ovld __cnfn popcount(ulong x);
-long2 __ovld __cnfn popcount(long2 x);
-ulong2 __ovld __cnfn popcount(ulong2 x);
-long3 __ovld __cnfn popcount(long3 x);
-ulong3 __ovld __cnfn popcount(ulong3 x);
-long4 __ovld __cnfn popcount(long4 x);
-ulong4 __ovld __cnfn popcount(ulong4 x);
-long8 __ovld __cnfn popcount(long8 x);
-ulong8 __ovld __cnfn popcount(ulong8 x);
-long16 __ovld __cnfn popcount(long16 x);
-ulong16 __ovld __cnfn popcount(ulong16 x);
-
-/**
- * Multiply two 24-bit integer values x and y and add
- * the 32-bit integer result to the 32-bit integer z.
- * Refer to definition of mul24 to see how the 24-bit
- * integer multiplication is performed.
- */
-int __ovld __cnfn mad24(int x, int y, int z);
-uint __ovld __cnfn mad24(uint x, uint y, uint z);
-int2 __ovld __cnfn mad24(int2 x, int2 y, int2 z);
-uint2 __ovld __cnfn mad24(uint2 x, uint2 y, uint2 z);
-int3 __ovld __cnfn mad24(int3 x, int3 y, int3 z);
-uint3 __ovld __cnfn mad24(uint3 x, uint3 y, uint3 z);
-int4 __ovld __cnfn mad24(int4 x, int4 y, int4 z);
-uint4 __ovld __cnfn mad24(uint4 x, uint4 y, uint4 z);
-int8 __ovld __cnfn mad24(int8 x, int8 y, int8 z);
-uint8 __ovld __cnfn mad24(uint8 x, uint8 y, uint8 z);
-int16 __ovld __cnfn mad24(int16 x, int16 y, int16 z);
-uint16 __ovld __cnfn mad24(uint16 x, uint16 y, uint16 z);
-
-/**
- * Multiply two 24-bit integer values x and y. x and y
- * are 32-bit integers but only the low 24-bits are used
- * to perform the multiplication. mul24 should only
- * be used when values in x and y are in the range [-
- * 2^23, 2^23-1] if x and y are signed integers and in the
- * range [0, 2^24-1] if x and y are unsigned integers. If
- * x and y are not in this range, the multiplication
- * result is implementation-defined.
- */
-int __ovld __cnfn mul24(int x, int y);
-uint __ovld __cnfn mul24(uint x, uint y);
-int2 __ovld __cnfn mul24(int2 x, int2 y);
-uint2 __ovld __cnfn mul24(uint2 x, uint2 y);
-int3 __ovld __cnfn mul24(int3 x, int3 y);
-uint3 __ovld __cnfn mul24(uint3 x, uint3 y);
-int4 __ovld __cnfn mul24(int4 x, int4 y);
-uint4 __ovld __cnfn mul24(uint4 x, uint4 y);
-int8 __ovld __cnfn mul24(int8 x, int8 y);
-uint8 __ovld __cnfn mul24(uint8 x, uint8 y);
-int16 __ovld __cnfn mul24(int16 x, int16 y);
-uint16 __ovld __cnfn mul24(uint16 x, uint16 y);
-
-// OpenCL v1.1 s6.11.4, v1.2 s6.12.4, v2.0 s6.13.4 - Common Functions
-
-/**
- * Returns fmin(fmax(x, minval), maxval).
- * Results are undefined if minval > maxval.
- */
-float __ovld __cnfn clamp(float x, float minval, float maxval);
-float2 __ovld __cnfn clamp(float2 x, float2 minval, float2 maxval);
-float3 __ovld __cnfn clamp(float3 x, float3 minval, float3 maxval);
-float4 __ovld __cnfn clamp(float4 x, float4 minval, float4 maxval);
-float8 __ovld __cnfn clamp(float8 x, float8 minval, float8 maxval);
-float16 __ovld __cnfn clamp(float16 x, float16 minval, float16 maxval);
-float2 __ovld __cnfn clamp(float2 x, float minval, float maxval);
-float3 __ovld __cnfn clamp(float3 x, float minval, float maxval);
-float4 __ovld __cnfn clamp(float4 x, float minval, float maxval);
-float8 __ovld __cnfn clamp(float8 x, float minval, float maxval);
-float16 __ovld __cnfn clamp(float16 x, float minval, float maxval);
-#ifdef cl_khr_fp64
-double __ovld __cnfn clamp(double x, double minval, double maxval);
-double2 __ovld __cnfn clamp(double2 x, double2 minval, double2 maxval);
-double3 __ovld __cnfn clamp(double3 x, double3 minval, double3 maxval);
-double4 __ovld __cnfn clamp(double4 x, double4 minval, double4 maxval);
-double8 __ovld __cnfn clamp(double8 x, double8 minval, double8 maxval);
-double16 __ovld __cnfn clamp(double16 x, double16 minval, double16 maxval);
-double2 __ovld __cnfn clamp(double2 x, double minval, double maxval);
-double3 __ovld __cnfn clamp(double3 x, double minval, double maxval);
-double4 __ovld __cnfn clamp(double4 x, double minval, double maxval);
-double8 __ovld __cnfn clamp(double8 x, double minval, double maxval);
-double16 __ovld __cnfn clamp(double16 x, double minval, double maxval);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn clamp(half x, half minval, half maxval);
-half2 __ovld __cnfn clamp(half2 x, half2 minval, half2 maxval);
-half3 __ovld __cnfn clamp(half3 x, half3 minval, half3 maxval);
-half4 __ovld __cnfn clamp(half4 x, half4 minval, half4 maxval);
-half8 __ovld __cnfn clamp(half8 x, half8 minval, half8 maxval);
-half16 __ovld __cnfn clamp(half16 x, half16 minval, half16 maxval);
-half2 __ovld __cnfn clamp(half2 x, half minval, half maxval);
-half3 __ovld __cnfn clamp(half3 x, half minval, half maxval);
-half4 __ovld __cnfn clamp(half4 x, half minval, half maxval);
-half8 __ovld __cnfn clamp(half8 x, half minval, half maxval);
-half16 __ovld __cnfn clamp(half16 x, half minval, half maxval);
-#endif //cl_khr_fp16
-
-/**
- * Converts radians to degrees, i.e. (180 / PI) *
- * radians.
- */
-float __ovld __cnfn degrees(float radians);
-float2 __ovld __cnfn degrees(float2 radians);
-float3 __ovld __cnfn degrees(float3 radians);
-float4 __ovld __cnfn degrees(float4 radians);
-float8 __ovld __cnfn degrees(float8 radians);
-float16 __ovld __cnfn degrees(float16 radians);
-#ifdef cl_khr_fp64
-double __ovld __cnfn degrees(double radians);
-double2 __ovld __cnfn degrees(double2 radians);
-double3 __ovld __cnfn degrees(double3 radians);
-double4 __ovld __cnfn degrees(double4 radians);
-double8 __ovld __cnfn degrees(double8 radians);
-double16 __ovld __cnfn degrees(double16 radians);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn degrees(half radians);
-half2 __ovld __cnfn degrees(half2 radians);
-half3 __ovld __cnfn degrees(half3 radians);
-half4 __ovld __cnfn degrees(half4 radians);
-half8 __ovld __cnfn degrees(half8 radians);
-half16 __ovld __cnfn degrees(half16 radians);
-#endif //cl_khr_fp16
-
-/**
- * Returns y if x < y, otherwise it returns x. If x and y
- * are infinite or NaN, the return values are undefined.
- */
-float __ovld __cnfn max(float x, float y);
-float2 __ovld __cnfn max(float2 x, float2 y);
-float3 __ovld __cnfn max(float3 x, float3 y);
-float4 __ovld __cnfn max(float4 x, float4 y);
-float8 __ovld __cnfn max(float8 x, float8 y);
-float16 __ovld __cnfn max(float16 x, float16 y);
-float2 __ovld __cnfn max(float2 x, float y);
-float3 __ovld __cnfn max(float3 x, float y);
-float4 __ovld __cnfn max(float4 x, float y);
-float8 __ovld __cnfn max(float8 x, float y);
-float16 __ovld __cnfn max(float16 x, float y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn max(double x, double y);
-double2 __ovld __cnfn max(double2 x, double2 y);
-double3 __ovld __cnfn max(double3 x, double3 y);
-double4 __ovld __cnfn max(double4 x, double4 y);
-double8 __ovld __cnfn max(double8 x, double8 y);
-double16 __ovld __cnfn max(double16 x, double16 y);
-double2 __ovld __cnfn max(double2 x, double y);
-double3 __ovld __cnfn max(double3 x, double y);
-double4 __ovld __cnfn max(double4 x, double y);
-double8 __ovld __cnfn max(double8 x, double y);
-double16 __ovld __cnfn max(double16 x, double y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn max(half x, half y);
-half2 __ovld __cnfn max(half2 x, half2 y);
-half3 __ovld __cnfn max(half3 x, half3 y);
-half4 __ovld __cnfn max(half4 x, half4 y);
-half8 __ovld __cnfn max(half8 x, half8 y);
-half16 __ovld __cnfn max(half16 x, half16 y);
-half2 __ovld __cnfn max(half2 x, half y);
-half3 __ovld __cnfn max(half3 x, half y);
-half4 __ovld __cnfn max(half4 x, half y);
-half8 __ovld __cnfn max(half8 x, half y);
-half16 __ovld __cnfn max(half16 x, half y);
-#endif //cl_khr_fp16
-
-/**
- * Returns y if y < x, otherwise it returns x. If x and y
- * are infinite or NaN, the return values are undefined.
- */
-float __ovld __cnfn min(float x, float y);
-float2 __ovld __cnfn min(float2 x, float2 y);
-float3 __ovld __cnfn min(float3 x, float3 y);
-float4 __ovld __cnfn min(float4 x, float4 y);
-float8 __ovld __cnfn min(float8 x, float8 y);
-float16 __ovld __cnfn min(float16 x, float16 y);
-float2 __ovld __cnfn min(float2 x, float y);
-float3 __ovld __cnfn min(float3 x, float y);
-float4 __ovld __cnfn min(float4 x, float y);
-float8 __ovld __cnfn min(float8 x, float y);
-float16 __ovld __cnfn min(float16 x, float y);
-#ifdef cl_khr_fp64
-double __ovld __cnfn min(double x, double y);
-double2 __ovld __cnfn min(double2 x, double2 y);
-double3 __ovld __cnfn min(double3 x, double3 y);
-double4 __ovld __cnfn min(double4 x, double4 y);
-double8 __ovld __cnfn min(double8 x, double8 y);
-double16 __ovld __cnfn min(double16 x, double16 y);
-double2 __ovld __cnfn min(double2 x, double y);
-double3 __ovld __cnfn min(double3 x, double y);
-double4 __ovld __cnfn min(double4 x, double y);
-double8 __ovld __cnfn min(double8 x, double y);
-double16 __ovld __cnfn min(double16 x, double y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn min(half x, half y);
-half2 __ovld __cnfn min(half2 x, half2 y);
-half3 __ovld __cnfn min(half3 x, half3 y);
-half4 __ovld __cnfn min(half4 x, half4 y);
-half8 __ovld __cnfn min(half8 x, half8 y);
-half16 __ovld __cnfn min(half16 x, half16 y);
-half2 __ovld __cnfn min(half2 x, half y);
-half3 __ovld __cnfn min(half3 x, half y);
-half4 __ovld __cnfn min(half4 x, half y);
-half8 __ovld __cnfn min(half8 x, half y);
-half16 __ovld __cnfn min(half16 x, half y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the linear blend of x & y implemented as:
- * x + (y - x) * a
- * a must be a value in the range 0.0 ... 1.0. If a is not
- * in the range 0.0 ... 1.0, the return values are
- * undefined.
- */
-float __ovld __cnfn mix(float x, float y, float a);
-float2 __ovld __cnfn mix(float2 x, float2 y, float2 a);
-float3 __ovld __cnfn mix(float3 x, float3 y, float3 a);
-float4 __ovld __cnfn mix(float4 x, float4 y, float4 a);
-float8 __ovld __cnfn mix(float8 x, float8 y, float8 a);
-float16 __ovld __cnfn mix(float16 x, float16 y, float16 a);
-float2 __ovld __cnfn mix(float2 x, float2 y, float a);
-float3 __ovld __cnfn mix(float3 x, float3 y, float a);
-float4 __ovld __cnfn mix(float4 x, float4 y, float a);
-float8 __ovld __cnfn mix(float8 x, float8 y, float a);
-float16 __ovld __cnfn mix(float16 x, float16 y, float a);
-#ifdef cl_khr_fp64
-double __ovld __cnfn mix(double x, double y, double a);
-double2 __ovld __cnfn mix(double2 x, double2 y, double2 a);
-double3 __ovld __cnfn mix(double3 x, double3 y, double3 a);
-double4 __ovld __cnfn mix(double4 x, double4 y, double4 a);
-double8 __ovld __cnfn mix(double8 x, double8 y, double8 a);
-double16 __ovld __cnfn mix(double16 x, double16 y, double16 a);
-double2 __ovld __cnfn mix(double2 x, double2 y, double a);
-double3 __ovld __cnfn mix(double3 x, double3 y, double a);
-double4 __ovld __cnfn mix(double4 x, double4 y, double a);
-double8 __ovld __cnfn mix(double8 x, double8 y, double a);
-double16 __ovld __cnfn mix(double16 x, double16 y, double a);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn mix(half x, half y, half a);
-half2 __ovld __cnfn mix(half2 x, half2 y, half2 a);
-half3 __ovld __cnfn mix(half3 x, half3 y, half3 a);
-half4 __ovld __cnfn mix(half4 x, half4 y, half4 a);
-half8 __ovld __cnfn mix(half8 x, half8 y, half8 a);
-half16 __ovld __cnfn mix(half16 x, half16 y, half16 a);
-half2 __ovld __cnfn mix(half2 x, half2 y, half a);
-half3 __ovld __cnfn mix(half3 x, half3 y, half a);
-half4 __ovld __cnfn mix(half4 x, half4 y, half a);
-half8 __ovld __cnfn mix(half8 x, half8 y, half a);
-half16 __ovld __cnfn mix(half16 x, half16 y, half a);
-#endif //cl_khr_fp16
-
-/**
- * Converts degrees to radians, i.e. (PI / 180) *
- * degrees.
- */
-float __ovld __cnfn radians(float degrees);
-float2 __ovld __cnfn radians(float2 degrees);
-float3 __ovld __cnfn radians(float3 degrees);
-float4 __ovld __cnfn radians(float4 degrees);
-float8 __ovld __cnfn radians(float8 degrees);
-float16 __ovld __cnfn radians(float16 degrees);
-#ifdef cl_khr_fp64
-double __ovld __cnfn radians(double degrees);
-double2 __ovld __cnfn radians(double2 degrees);
-double3 __ovld __cnfn radians(double3 degrees);
-double4 __ovld __cnfn radians(double4 degrees);
-double8 __ovld __cnfn radians(double8 degrees);
-double16 __ovld __cnfn radians(double16 degrees);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn radians(half degrees);
-half2 __ovld __cnfn radians(half2 degrees);
-half3 __ovld __cnfn radians(half3 degrees);
-half4 __ovld __cnfn radians(half4 degrees);
-half8 __ovld __cnfn radians(half8 degrees);
-half16 __ovld __cnfn radians(half16 degrees);
-#endif //cl_khr_fp16
-
-/**
- * Returns 0.0 if x < edge, otherwise it returns 1.0.
- */
-float __ovld __cnfn step(float edge, float x);
-float2 __ovld __cnfn step(float2 edge, float2 x);
-float3 __ovld __cnfn step(float3 edge, float3 x);
-float4 __ovld __cnfn step(float4 edge, float4 x);
-float8 __ovld __cnfn step(float8 edge, float8 x);
-float16 __ovld __cnfn step(float16 edge, float16 x);
-float2 __ovld __cnfn step(float edge, float2 x);
-float3 __ovld __cnfn step(float edge, float3 x);
-float4 __ovld __cnfn step(float edge, float4 x);
-float8 __ovld __cnfn step(float edge, float8 x);
-float16 __ovld __cnfn step(float edge, float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn step(double edge, double x);
-double2 __ovld __cnfn step(double2 edge, double2 x);
-double3 __ovld __cnfn step(double3 edge, double3 x);
-double4 __ovld __cnfn step(double4 edge, double4 x);
-double8 __ovld __cnfn step(double8 edge, double8 x);
-double16 __ovld __cnfn step(double16 edge, double16 x);
-double2 __ovld __cnfn step(double edge, double2 x);
-double3 __ovld __cnfn step(double edge, double3 x);
-double4 __ovld __cnfn step(double edge, double4 x);
-double8 __ovld __cnfn step(double edge, double8 x);
-double16 __ovld __cnfn step(double edge, double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn step(half edge, half x);
-half2 __ovld __cnfn step(half2 edge, half2 x);
-half3 __ovld __cnfn step(half3 edge, half3 x);
-half4 __ovld __cnfn step(half4 edge, half4 x);
-half8 __ovld __cnfn step(half8 edge, half8 x);
-half16 __ovld __cnfn step(half16 edge, half16 x);
-half2 __ovld __cnfn step(half edge, half2 x);
-half3 __ovld __cnfn step(half edge, half3 x);
-half4 __ovld __cnfn step(half edge, half4 x);
-half8 __ovld __cnfn step(half edge, half8 x);
-half16 __ovld __cnfn step(half edge, half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Returns 0.0 if x <= edge0 and 1.0 if x >= edge1 and
- * performs smooth Hermite interpolation between 0
- * and 1when edge0 < x < edge1. This is useful in
- * cases where you would want a threshold function
- * with a smooth transition.
- * This is equivalent to:
- * gentype t;
- * t = clamp ((x - edge0) / (edge1 - edge0), 0, 1);
- * return t * t * (3 - 2 * t);
- * Results are undefined if edge0 >= edge1 or if x,
- * edge0 or edge1 is a NaN.
- */
-float __ovld __cnfn smoothstep(float edge0, float edge1, float x);
-float2 __ovld __cnfn smoothstep(float2 edge0, float2 edge1, float2 x);
-float3 __ovld __cnfn smoothstep(float3 edge0, float3 edge1, float3 x);
-float4 __ovld __cnfn smoothstep(float4 edge0, float4 edge1, float4 x);
-float8 __ovld __cnfn smoothstep(float8 edge0, float8 edge1, float8 x);
-float16 __ovld __cnfn smoothstep(float16 edge0, float16 edge1, float16 x);
-float2 __ovld __cnfn smoothstep(float edge0, float edge1, float2 x);
-float3 __ovld __cnfn smoothstep(float edge0, float edge1, float3 x);
-float4 __ovld __cnfn smoothstep(float edge0, float edge1, float4 x);
-float8 __ovld __cnfn smoothstep(float edge0, float edge1, float8 x);
-float16 __ovld __cnfn smoothstep(float edge0, float edge1, float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn smoothstep(double edge0, double edge1, double x);
-double2 __ovld __cnfn smoothstep(double2 edge0, double2 edge1, double2 x);
-double3 __ovld __cnfn smoothstep(double3 edge0, double3 edge1, double3 x);
-double4 __ovld __cnfn smoothstep(double4 edge0, double4 edge1, double4 x);
-double8 __ovld __cnfn smoothstep(double8 edge0, double8 edge1, double8 x);
-double16 __ovld __cnfn smoothstep(double16 edge0, double16 edge1, double16 x);
-double2 __ovld __cnfn smoothstep(double edge0, double edge1, double2 x);
-double3 __ovld __cnfn smoothstep(double edge0, double edge1, double3 x);
-double4 __ovld __cnfn smoothstep(double edge0, double edge1, double4 x);
-double8 __ovld __cnfn smoothstep(double edge0, double edge1, double8 x);
-double16 __ovld __cnfn smoothstep(double edge0, double edge1, double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn smoothstep(half edge0, half edge1, half x);
-half2 __ovld __cnfn smoothstep(half2 edge0, half2 edge1, half2 x);
-half3 __ovld __cnfn smoothstep(half3 edge0, half3 edge1, half3 x);
-half4 __ovld __cnfn smoothstep(half4 edge0, half4 edge1, half4 x);
-half8 __ovld __cnfn smoothstep(half8 edge0, half8 edge1, half8 x);
-half16 __ovld __cnfn smoothstep(half16 edge0, half16 edge1, half16 x);
-half2 __ovld __cnfn smoothstep(half edge0, half edge1, half2 x);
-half3 __ovld __cnfn smoothstep(half edge0, half edge1, half3 x);
-half4 __ovld __cnfn smoothstep(half edge0, half edge1, half4 x);
-half8 __ovld __cnfn smoothstep(half edge0, half edge1, half8 x);
-half16 __ovld __cnfn smoothstep(half edge0, half edge1, half16 x);
-#endif //cl_khr_fp16
-
-/**
- * Returns 1.0 if x > 0, -0.0 if x = -0.0, +0.0 if x =
- * +0.0, or -1.0 if x < 0. Returns 0.0 if x is a NaN.
- */
-float __ovld __cnfn sign(float x);
-float2 __ovld __cnfn sign(float2 x);
-float3 __ovld __cnfn sign(float3 x);
-float4 __ovld __cnfn sign(float4 x);
-float8 __ovld __cnfn sign(float8 x);
-float16 __ovld __cnfn sign(float16 x);
-#ifdef cl_khr_fp64
-double __ovld __cnfn sign(double x);
-double2 __ovld __cnfn sign(double2 x);
-double3 __ovld __cnfn sign(double3 x);
-double4 __ovld __cnfn sign(double4 x);
-double8 __ovld __cnfn sign(double8 x);
-double16 __ovld __cnfn sign(double16 x);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn sign(half x);
-half2 __ovld __cnfn sign(half2 x);
-half3 __ovld __cnfn sign(half3 x);
-half4 __ovld __cnfn sign(half4 x);
-half8 __ovld __cnfn sign(half8 x);
-half16 __ovld __cnfn sign(half16 x);
-#endif //cl_khr_fp16
-
-// OpenCL v1.1 s6.11.5, v1.2 s6.12.5, v2.0 s6.13.5 - Geometric Functions
-
-/**
- * Returns the cross product of p0.xyz and p1.xyz. The
- * w component of float4 result returned will be 0.0.
- */
-float4 __ovld __cnfn cross(float4 p0, float4 p1);
-float3 __ovld __cnfn cross(float3 p0, float3 p1);
-#ifdef cl_khr_fp64
-double4 __ovld __cnfn cross(double4 p0, double4 p1);
-double3 __ovld __cnfn cross(double3 p0, double3 p1);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half4 __ovld __cnfn cross(half4 p0, half4 p1);
-half3 __ovld __cnfn cross(half3 p0, half3 p1);
-#endif //cl_khr_fp16
-
-/**
- * Compute dot product.
- */
-float __ovld __cnfn dot(float p0, float p1);
-float __ovld __cnfn dot(float2 p0, float2 p1);
-float __ovld __cnfn dot(float3 p0, float3 p1);
-float __ovld __cnfn dot(float4 p0, float4 p1);
-#ifdef cl_khr_fp64
-double __ovld __cnfn dot(double p0, double p1);
-double __ovld __cnfn dot(double2 p0, double2 p1);
-double __ovld __cnfn dot(double3 p0, double3 p1);
-double __ovld __cnfn dot(double4 p0, double4 p1);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn dot(half p0, half p1);
-half __ovld __cnfn dot(half2 p0, half2 p1);
-half __ovld __cnfn dot(half3 p0, half3 p1);
-half __ovld __cnfn dot(half4 p0, half4 p1);
-#endif //cl_khr_fp16
-
-/**
- * Returns the distance between p0 and p1. This is
- * calculated as length(p0 - p1).
- */
-float __ovld __cnfn distance(float p0, float p1);
-float __ovld __cnfn distance(float2 p0, float2 p1);
-float __ovld __cnfn distance(float3 p0, float3 p1);
-float __ovld __cnfn distance(float4 p0, float4 p1);
-#ifdef cl_khr_fp64
-double __ovld __cnfn distance(double p0, double p1);
-double __ovld __cnfn distance(double2 p0, double2 p1);
-double __ovld __cnfn distance(double3 p0, double3 p1);
-double __ovld __cnfn distance(double4 p0, double4 p1);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn distance(half p0, half p1);
-half __ovld __cnfn distance(half2 p0, half2 p1);
-half __ovld __cnfn distance(half3 p0, half3 p1);
-half __ovld __cnfn distance(half4 p0, half4 p1);
-#endif //cl_khr_fp16
-
-/**
- * Return the length of vector p, i.e.,
- * sqrt(p.x2 + p.y 2 + ...)
- */
-float __ovld __cnfn length(float p);
-float __ovld __cnfn length(float2 p);
-float __ovld __cnfn length(float3 p);
-float __ovld __cnfn length(float4 p);
-#ifdef cl_khr_fp64
-double __ovld __cnfn length(double p);
-double __ovld __cnfn length(double2 p);
-double __ovld __cnfn length(double3 p);
-double __ovld __cnfn length(double4 p);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn length(half p);
-half __ovld __cnfn length(half2 p);
-half __ovld __cnfn length(half3 p);
-half __ovld __cnfn length(half4 p);
-#endif //cl_khr_fp16
-
-/**
- * Returns a vector in the same direction as p but with a
- * length of 1.
- */
-float __ovld __cnfn normalize(float p);
-float2 __ovld __cnfn normalize(float2 p);
-float3 __ovld __cnfn normalize(float3 p);
-float4 __ovld __cnfn normalize(float4 p);
-#ifdef cl_khr_fp64
-double __ovld __cnfn normalize(double p);
-double2 __ovld __cnfn normalize(double2 p);
-double3 __ovld __cnfn normalize(double3 p);
-double4 __ovld __cnfn normalize(double4 p);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn normalize(half p);
-half2 __ovld __cnfn normalize(half2 p);
-half3 __ovld __cnfn normalize(half3 p);
-half4 __ovld __cnfn normalize(half4 p);
-#endif //cl_khr_fp16
-
-/**
- * Returns fast_length(p0 - p1).
- */
-float __ovld __cnfn fast_distance(float p0, float p1);
-float __ovld __cnfn fast_distance(float2 p0, float2 p1);
-float __ovld __cnfn fast_distance(float3 p0, float3 p1);
-float __ovld __cnfn fast_distance(float4 p0, float4 p1);
-#ifdef cl_khr_fp16
-half __ovld __cnfn fast_distance(half p0, half p1);
-half __ovld __cnfn fast_distance(half2 p0, half2 p1);
-half __ovld __cnfn fast_distance(half3 p0, half3 p1);
-half __ovld __cnfn fast_distance(half4 p0, half4 p1);
-#endif //cl_khr_fp16
-
-/**
- * Returns the length of vector p computed as:
- * half_sqrt(p.x2 + p.y2 + ...)
- */
-float __ovld __cnfn fast_length(float p);
-float __ovld __cnfn fast_length(float2 p);
-float __ovld __cnfn fast_length(float3 p);
-float __ovld __cnfn fast_length(float4 p);
-#ifdef cl_khr_fp16
-half __ovld __cnfn fast_length(half p);
-half __ovld __cnfn fast_length(half2 p);
-half __ovld __cnfn fast_length(half3 p);
-half __ovld __cnfn fast_length(half4 p);
-#endif //cl_khr_fp16
-
-/**
- * Returns a vector in the same direction as p but with a
- * length of 1. fast_normalize is computed as:
- * p * half_rsqrt (p.x^2 + p.y^2 + ... )
- * The result shall be within 8192 ulps error from the
- * infinitely precise result of
- * if (all(p == 0.0f))
- * result = p;
- * else
- * result = p / sqrt (p.x^2 + p.y^2 + ...);
- * with the following exceptions:
- * 1) If the sum of squares is greater than FLT_MAX
- * then the value of the floating-point values in the
- * result vector are undefined.
- * 2) If the sum of squares is less than FLT_MIN then
- * the implementation may return back p.
- * 3) If the device is in "denorms are flushed to zero"
- * mode, individual operand elements with magnitude
- * less than sqrt(FLT_MIN) may be flushed to zero
- * before proceeding with the calculation.
- */
-float __ovld __cnfn fast_normalize(float p);
-float2 __ovld __cnfn fast_normalize(float2 p);
-float3 __ovld __cnfn fast_normalize(float3 p);
-float4 __ovld __cnfn fast_normalize(float4 p);
-#ifdef cl_khr_fp16
-half __ovld __cnfn fast_normalize(half p);
-half2 __ovld __cnfn fast_normalize(half2 p);
-half3 __ovld __cnfn fast_normalize(half3 p);
-half4 __ovld __cnfn fast_normalize(half4 p);
-#endif //cl_khr_fp16
-
-// OpenCL v1.1 s6.11.6, v1.2 s6.12.6, v2.0 s6.13.6 - Relational Functions
-
-/**
- * intn isequal (floatn x, floatn y)
- * Returns the component-wise compare of x == y.
- */
-int __ovld __cnfn isequal(float x, float y);
-int2 __ovld __cnfn isequal(float2 x, float2 y);
-int3 __ovld __cnfn isequal(float3 x, float3 y);
-int4 __ovld __cnfn isequal(float4 x, float4 y);
-int8 __ovld __cnfn isequal(float8 x, float8 y);
-int16 __ovld __cnfn isequal(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isequal(double x, double y);
-long2 __ovld __cnfn isequal(double2 x, double2 y);
-long3 __ovld __cnfn isequal(double3 x, double3 y);
-long4 __ovld __cnfn isequal(double4 x, double4 y);
-long8 __ovld __cnfn isequal(double8 x, double8 y);
-long16 __ovld __cnfn isequal(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isequal(half x, half y);
-short2 __ovld __cnfn isequal(half2 x, half2 y);
-short3 __ovld __cnfn isequal(half3 x, half3 y);
-short4 __ovld __cnfn isequal(half4 x, half4 y);
-short8 __ovld __cnfn isequal(half8 x, half8 y);
-short16 __ovld __cnfn isequal(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of x != y.
- */
-int __ovld __cnfn isnotequal(float x, float y);
-int2 __ovld __cnfn isnotequal(float2 x, float2 y);
-int3 __ovld __cnfn isnotequal(float3 x, float3 y);
-int4 __ovld __cnfn isnotequal(float4 x, float4 y);
-int8 __ovld __cnfn isnotequal(float8 x, float8 y);
-int16 __ovld __cnfn isnotequal(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isnotequal(double x, double y);
-long2 __ovld __cnfn isnotequal(double2 x, double2 y);
-long3 __ovld __cnfn isnotequal(double3 x, double3 y);
-long4 __ovld __cnfn isnotequal(double4 x, double4 y);
-long8 __ovld __cnfn isnotequal(double8 x, double8 y);
-long16 __ovld __cnfn isnotequal(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isnotequal(half x, half y);
-short2 __ovld __cnfn isnotequal(half2 x, half2 y);
-short3 __ovld __cnfn isnotequal(half3 x, half3 y);
-short4 __ovld __cnfn isnotequal(half4 x, half4 y);
-short8 __ovld __cnfn isnotequal(half8 x, half8 y);
-short16 __ovld __cnfn isnotequal(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of x > y.
- */
-int __ovld __cnfn isgreater(float x, float y);
-int2 __ovld __cnfn isgreater(float2 x, float2 y);
-int3 __ovld __cnfn isgreater(float3 x, float3 y);
-int4 __ovld __cnfn isgreater(float4 x, float4 y);
-int8 __ovld __cnfn isgreater(float8 x, float8 y);
-int16 __ovld __cnfn isgreater(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isgreater(double x, double y);
-long2 __ovld __cnfn isgreater(double2 x, double2 y);
-long3 __ovld __cnfn isgreater(double3 x, double3 y);
-long4 __ovld __cnfn isgreater(double4 x, double4 y);
-long8 __ovld __cnfn isgreater(double8 x, double8 y);
-long16 __ovld __cnfn isgreater(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isgreater(half x, half y);
-short2 __ovld __cnfn isgreater(half2 x, half2 y);
-short3 __ovld __cnfn isgreater(half3 x, half3 y);
-short4 __ovld __cnfn isgreater(half4 x, half4 y);
-short8 __ovld __cnfn isgreater(half8 x, half8 y);
-short16 __ovld __cnfn isgreater(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of x >= y.
- */
-int __ovld __cnfn isgreaterequal(float x, float y);
-int2 __ovld __cnfn isgreaterequal(float2 x, float2 y);
-int3 __ovld __cnfn isgreaterequal(float3 x, float3 y);
-int4 __ovld __cnfn isgreaterequal(float4 x, float4 y);
-int8 __ovld __cnfn isgreaterequal(float8 x, float8 y);
-int16 __ovld __cnfn isgreaterequal(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isgreaterequal(double x, double y);
-long2 __ovld __cnfn isgreaterequal(double2 x, double2 y);
-long3 __ovld __cnfn isgreaterequal(double3 x, double3 y);
-long4 __ovld __cnfn isgreaterequal(double4 x, double4 y);
-long8 __ovld __cnfn isgreaterequal(double8 x, double8 y);
-long16 __ovld __cnfn isgreaterequal(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isgreaterequal(half x, half y);
-short2 __ovld __cnfn isgreaterequal(half2 x, half2 y);
-short3 __ovld __cnfn isgreaterequal(half3 x, half3 y);
-short4 __ovld __cnfn isgreaterequal(half4 x, half4 y);
-short8 __ovld __cnfn isgreaterequal(half8 x, half8 y);
-short16 __ovld __cnfn isgreaterequal(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of x < y.
- */
-int __ovld __cnfn isless(float x, float y);
-int2 __ovld __cnfn isless(float2 x, float2 y);
-int3 __ovld __cnfn isless(float3 x, float3 y);
-int4 __ovld __cnfn isless(float4 x, float4 y);
-int8 __ovld __cnfn isless(float8 x, float8 y);
-int16 __ovld __cnfn isless(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isless(double x, double y);
-long2 __ovld __cnfn isless(double2 x, double2 y);
-long3 __ovld __cnfn isless(double3 x, double3 y);
-long4 __ovld __cnfn isless(double4 x, double4 y);
-long8 __ovld __cnfn isless(double8 x, double8 y);
-long16 __ovld __cnfn isless(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isless(half x, half y);
-short2 __ovld __cnfn isless(half2 x, half2 y);
-short3 __ovld __cnfn isless(half3 x, half3 y);
-short4 __ovld __cnfn isless(half4 x, half4 y);
-short8 __ovld __cnfn isless(half8 x, half8 y);
-short16 __ovld __cnfn isless(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of x <= y.
- */
-int __ovld __cnfn islessequal(float x, float y);
-int2 __ovld __cnfn islessequal(float2 x, float2 y);
-int3 __ovld __cnfn islessequal(float3 x, float3 y);
-int4 __ovld __cnfn islessequal(float4 x, float4 y);
-int8 __ovld __cnfn islessequal(float8 x, float8 y);
-int16 __ovld __cnfn islessequal(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn islessequal(double x, double y);
-long2 __ovld __cnfn islessequal(double2 x, double2 y);
-long3 __ovld __cnfn islessequal(double3 x, double3 y);
-long4 __ovld __cnfn islessequal(double4 x, double4 y);
-long8 __ovld __cnfn islessequal(double8 x, double8 y);
-long16 __ovld __cnfn islessequal(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn islessequal(half x, half y);
-short2 __ovld __cnfn islessequal(half2 x, half2 y);
-short3 __ovld __cnfn islessequal(half3 x, half3 y);
-short4 __ovld __cnfn islessequal(half4 x, half4 y);
-short8 __ovld __cnfn islessequal(half8 x, half8 y);
-short16 __ovld __cnfn islessequal(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Returns the component-wise compare of
- * (x < y) || (x > y) .
- */
-int __ovld __cnfn islessgreater(float x, float y);
-int2 __ovld __cnfn islessgreater(float2 x, float2 y);
-int3 __ovld __cnfn islessgreater(float3 x, float3 y);
-int4 __ovld __cnfn islessgreater(float4 x, float4 y);
-int8 __ovld __cnfn islessgreater(float8 x, float8 y);
-int16 __ovld __cnfn islessgreater(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn islessgreater(double x, double y);
-long2 __ovld __cnfn islessgreater(double2 x, double2 y);
-long3 __ovld __cnfn islessgreater(double3 x, double3 y);
-long4 __ovld __cnfn islessgreater(double4 x, double4 y);
-long8 __ovld __cnfn islessgreater(double8 x, double8 y);
-long16 __ovld __cnfn islessgreater(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn islessgreater(half x, half y);
-short2 __ovld __cnfn islessgreater(half2 x, half2 y);
-short3 __ovld __cnfn islessgreater(half3 x, half3 y);
-short4 __ovld __cnfn islessgreater(half4 x, half4 y);
-short8 __ovld __cnfn islessgreater(half8 x, half8 y);
-short16 __ovld __cnfn islessgreater(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Test for finite value.
- */
-int __ovld __cnfn isfinite(float);
-int2 __ovld __cnfn isfinite(float2);
-int3 __ovld __cnfn isfinite(float3);
-int4 __ovld __cnfn isfinite(float4);
-int8 __ovld __cnfn isfinite(float8);
-int16 __ovld __cnfn isfinite(float16);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isfinite(double);
-long2 __ovld __cnfn isfinite(double2);
-long3 __ovld __cnfn isfinite(double3);
-long4 __ovld __cnfn isfinite(double4);
-long8 __ovld __cnfn isfinite(double8);
-long16 __ovld __cnfn isfinite(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isfinite(half);
-short2 __ovld __cnfn isfinite(half2);
-short3 __ovld __cnfn isfinite(half3);
-short4 __ovld __cnfn isfinite(half4);
-short8 __ovld __cnfn isfinite(half8);
-short16 __ovld __cnfn isfinite(half16);
-#endif //cl_khr_fp16
-
-/**
- * Test for infinity value (+ve or -ve) .
- */
-int __ovld __cnfn isinf(float);
-int2 __ovld __cnfn isinf(float2);
-int3 __ovld __cnfn isinf(float3);
-int4 __ovld __cnfn isinf(float4);
-int8 __ovld __cnfn isinf(float8);
-int16 __ovld __cnfn isinf(float16);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isinf(double);
-long2 __ovld __cnfn isinf(double2);
-long3 __ovld __cnfn isinf(double3);
-long4 __ovld __cnfn isinf(double4);
-long8 __ovld __cnfn isinf(double8);
-long16 __ovld __cnfn isinf(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isinf(half);
-short2 __ovld __cnfn isinf(half2);
-short3 __ovld __cnfn isinf(half3);
-short4 __ovld __cnfn isinf(half4);
-short8 __ovld __cnfn isinf(half8);
-short16 __ovld __cnfn isinf(half16);
-#endif //cl_khr_fp16
-
-/**
- * Test for a NaN.
- */
-int __ovld __cnfn isnan(float);
-int2 __ovld __cnfn isnan(float2);
-int3 __ovld __cnfn isnan(float3);
-int4 __ovld __cnfn isnan(float4);
-int8 __ovld __cnfn isnan(float8);
-int16 __ovld __cnfn isnan(float16);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isnan(double);
-long2 __ovld __cnfn isnan(double2);
-long3 __ovld __cnfn isnan(double3);
-long4 __ovld __cnfn isnan(double4);
-long8 __ovld __cnfn isnan(double8);
-long16 __ovld __cnfn isnan(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isnan(half);
-short2 __ovld __cnfn isnan(half2);
-short3 __ovld __cnfn isnan(half3);
-short4 __ovld __cnfn isnan(half4);
-short8 __ovld __cnfn isnan(half8);
-short16 __ovld __cnfn isnan(half16);
-#endif //cl_khr_fp16
-
-/**
- * Test for a normal value.
- */
-int __ovld __cnfn isnormal(float);
-int2 __ovld __cnfn isnormal(float2);
-int3 __ovld __cnfn isnormal(float3);
-int4 __ovld __cnfn isnormal(float4);
-int8 __ovld __cnfn isnormal(float8);
-int16 __ovld __cnfn isnormal(float16);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isnormal(double);
-long2 __ovld __cnfn isnormal(double2);
-long3 __ovld __cnfn isnormal(double3);
-long4 __ovld __cnfn isnormal(double4);
-long8 __ovld __cnfn isnormal(double8);
-long16 __ovld __cnfn isnormal(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isnormal(half);
-short2 __ovld __cnfn isnormal(half2);
-short3 __ovld __cnfn isnormal(half3);
-short4 __ovld __cnfn isnormal(half4);
-short8 __ovld __cnfn isnormal(half8);
-short16 __ovld __cnfn isnormal(half16);
-#endif //cl_khr_fp16
-
-/**
- * Test if arguments are ordered. isordered() takes
- * arguments x and y, and returns the result
- * isequal(x, x) && isequal(y, y).
- */
-int __ovld __cnfn isordered(float x, float y);
-int2 __ovld __cnfn isordered(float2 x, float2 y);
-int3 __ovld __cnfn isordered(float3 x, float3 y);
-int4 __ovld __cnfn isordered(float4 x, float4 y);
-int8 __ovld __cnfn isordered(float8 x, float8 y);
-int16 __ovld __cnfn isordered(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isordered(double x, double y);
-long2 __ovld __cnfn isordered(double2 x, double2 y);
-long3 __ovld __cnfn isordered(double3 x, double3 y);
-long4 __ovld __cnfn isordered(double4 x, double4 y);
-long8 __ovld __cnfn isordered(double8 x, double8 y);
-long16 __ovld __cnfn isordered(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isordered(half x, half y);
-short2 __ovld __cnfn isordered(half2 x, half2 y);
-short3 __ovld __cnfn isordered(half3 x, half3 y);
-short4 __ovld __cnfn isordered(half4 x, half4 y);
-short8 __ovld __cnfn isordered(half8 x, half8 y);
-short16 __ovld __cnfn isordered(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Test if arguments are unordered. isunordered()
- * takes arguments x and y, returning non-zero if x or y
- * is NaN, and zero otherwise.
- */
-int __ovld __cnfn isunordered(float x, float y);
-int2 __ovld __cnfn isunordered(float2 x, float2 y);
-int3 __ovld __cnfn isunordered(float3 x, float3 y);
-int4 __ovld __cnfn isunordered(float4 x, float4 y);
-int8 __ovld __cnfn isunordered(float8 x, float8 y);
-int16 __ovld __cnfn isunordered(float16 x, float16 y);
-#ifdef cl_khr_fp64
-int __ovld __cnfn isunordered(double x, double y);
-long2 __ovld __cnfn isunordered(double2 x, double2 y);
-long3 __ovld __cnfn isunordered(double3 x, double3 y);
-long4 __ovld __cnfn isunordered(double4 x, double4 y);
-long8 __ovld __cnfn isunordered(double8 x, double8 y);
-long16 __ovld __cnfn isunordered(double16 x, double16 y);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn isunordered(half x, half y);
-short2 __ovld __cnfn isunordered(half2 x, half2 y);
-short3 __ovld __cnfn isunordered(half3 x, half3 y);
-short4 __ovld __cnfn isunordered(half4 x, half4 y);
-short8 __ovld __cnfn isunordered(half8 x, half8 y);
-short16 __ovld __cnfn isunordered(half16 x, half16 y);
-#endif //cl_khr_fp16
-
-/**
- * Test for sign bit. The scalar version of the function
- * returns a 1 if the sign bit in the float is set else returns
- * 0. The vector version of the function returns the
- * following for each component in floatn: a -1 if the
- * sign bit in the float is set else returns 0.
- */
-int __ovld __cnfn signbit(float);
-int2 __ovld __cnfn signbit(float2);
-int3 __ovld __cnfn signbit(float3);
-int4 __ovld __cnfn signbit(float4);
-int8 __ovld __cnfn signbit(float8);
-int16 __ovld __cnfn signbit(float16);
-#ifdef cl_khr_fp64
-int __ovld __cnfn signbit(double);
-long2 __ovld __cnfn signbit(double2);
-long3 __ovld __cnfn signbit(double3);
-long4 __ovld __cnfn signbit(double4);
-long8 __ovld __cnfn signbit(double8);
-long16 __ovld __cnfn signbit(double16);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-int __ovld __cnfn signbit(half);
-short2 __ovld __cnfn signbit(half2);
-short3 __ovld __cnfn signbit(half3);
-short4 __ovld __cnfn signbit(half4);
-short8 __ovld __cnfn signbit(half8);
-short16 __ovld __cnfn signbit(half16);
-#endif //cl_khr_fp16
-
-/**
- * Returns 1 if the most significant bit in any component
- * of x is set; otherwise returns 0.
- */
-int __ovld __cnfn any(char x);
-int __ovld __cnfn any(char2 x);
-int __ovld __cnfn any(char3 x);
-int __ovld __cnfn any(char4 x);
-int __ovld __cnfn any(char8 x);
-int __ovld __cnfn any(char16 x);
-int __ovld __cnfn any(short x);
-int __ovld __cnfn any(short2 x);
-int __ovld __cnfn any(short3 x);
-int __ovld __cnfn any(short4 x);
-int __ovld __cnfn any(short8 x);
-int __ovld __cnfn any(short16 x);
-int __ovld __cnfn any(int x);
-int __ovld __cnfn any(int2 x);
-int __ovld __cnfn any(int3 x);
-int __ovld __cnfn any(int4 x);
-int __ovld __cnfn any(int8 x);
-int __ovld __cnfn any(int16 x);
-int __ovld __cnfn any(long x);
-int __ovld __cnfn any(long2 x);
-int __ovld __cnfn any(long3 x);
-int __ovld __cnfn any(long4 x);
-int __ovld __cnfn any(long8 x);
-int __ovld __cnfn any(long16 x);
-
-/**
- * Returns 1 if the most significant bit in all components
- * of x is set; otherwise returns 0.
- */
-int __ovld __cnfn all(char x);
-int __ovld __cnfn all(char2 x);
-int __ovld __cnfn all(char3 x);
-int __ovld __cnfn all(char4 x);
-int __ovld __cnfn all(char8 x);
-int __ovld __cnfn all(char16 x);
-int __ovld __cnfn all(short x);
-int __ovld __cnfn all(short2 x);
-int __ovld __cnfn all(short3 x);
-int __ovld __cnfn all(short4 x);
-int __ovld __cnfn all(short8 x);
-int __ovld __cnfn all(short16 x);
-int __ovld __cnfn all(int x);
-int __ovld __cnfn all(int2 x);
-int __ovld __cnfn all(int3 x);
-int __ovld __cnfn all(int4 x);
-int __ovld __cnfn all(int8 x);
-int __ovld __cnfn all(int16 x);
-int __ovld __cnfn all(long x);
-int __ovld __cnfn all(long2 x);
-int __ovld __cnfn all(long3 x);
-int __ovld __cnfn all(long4 x);
-int __ovld __cnfn all(long8 x);
-int __ovld __cnfn all(long16 x);
-
-/**
- * Each bit of the result is the corresponding bit of a if
- * the corresponding bit of c is 0. Otherwise it is the
- * corresponding bit of b.
- */
-char __ovld __cnfn bitselect(char a, char b, char c);
-uchar __ovld __cnfn bitselect(uchar a, uchar b, uchar c);
-char2 __ovld __cnfn bitselect(char2 a, char2 b, char2 c);
-uchar2 __ovld __cnfn bitselect(uchar2 a, uchar2 b, uchar2 c);
-char3 __ovld __cnfn bitselect(char3 a, char3 b, char3 c);
-uchar3 __ovld __cnfn bitselect(uchar3 a, uchar3 b, uchar3 c);
-char4 __ovld __cnfn bitselect(char4 a, char4 b, char4 c);
-uchar4 __ovld __cnfn bitselect(uchar4 a, uchar4 b, uchar4 c);
-char8 __ovld __cnfn bitselect(char8 a, char8 b, char8 c);
-uchar8 __ovld __cnfn bitselect(uchar8 a, uchar8 b, uchar8 c);
-char16 __ovld __cnfn bitselect(char16 a, char16 b, char16 c);
-uchar16 __ovld __cnfn bitselect(uchar16 a, uchar16 b, uchar16 c);
-short __ovld __cnfn bitselect(short a, short b, short c);
-ushort __ovld __cnfn bitselect(ushort a, ushort b, ushort c);
-short2 __ovld __cnfn bitselect(short2 a, short2 b, short2 c);
-ushort2 __ovld __cnfn bitselect(ushort2 a, ushort2 b, ushort2 c);
-short3 __ovld __cnfn bitselect(short3 a, short3 b, short3 c);
-ushort3 __ovld __cnfn bitselect(ushort3 a, ushort3 b, ushort3 c);
-short4 __ovld __cnfn bitselect(short4 a, short4 b, short4 c);
-ushort4 __ovld __cnfn bitselect(ushort4 a, ushort4 b, ushort4 c);
-short8 __ovld __cnfn bitselect(short8 a, short8 b, short8 c);
-ushort8 __ovld __cnfn bitselect(ushort8 a, ushort8 b, ushort8 c);
-short16 __ovld __cnfn bitselect(short16 a, short16 b, short16 c);
-ushort16 __ovld __cnfn bitselect(ushort16 a, ushort16 b, ushort16 c);
-int __ovld __cnfn bitselect(int a, int b, int c);
-uint __ovld __cnfn bitselect(uint a, uint b, uint c);
-int2 __ovld __cnfn bitselect(int2 a, int2 b, int2 c);
-uint2 __ovld __cnfn bitselect(uint2 a, uint2 b, uint2 c);
-int3 __ovld __cnfn bitselect(int3 a, int3 b, int3 c);
-uint3 __ovld __cnfn bitselect(uint3 a, uint3 b, uint3 c);
-int4 __ovld __cnfn bitselect(int4 a, int4 b, int4 c);
-uint4 __ovld __cnfn bitselect(uint4 a, uint4 b, uint4 c);
-int8 __ovld __cnfn bitselect(int8 a, int8 b, int8 c);
-uint8 __ovld __cnfn bitselect(uint8 a, uint8 b, uint8 c);
-int16 __ovld __cnfn bitselect(int16 a, int16 b, int16 c);
-uint16 __ovld __cnfn bitselect(uint16 a, uint16 b, uint16 c);
-long __ovld __cnfn bitselect(long a, long b, long c);
-ulong __ovld __cnfn bitselect(ulong a, ulong b, ulong c);
-long2 __ovld __cnfn bitselect(long2 a, long2 b, long2 c);
-ulong2 __ovld __cnfn bitselect(ulong2 a, ulong2 b, ulong2 c);
-long3 __ovld __cnfn bitselect(long3 a, long3 b, long3 c);
-ulong3 __ovld __cnfn bitselect(ulong3 a, ulong3 b, ulong3 c);
-long4 __ovld __cnfn bitselect(long4 a, long4 b, long4 c);
-ulong4 __ovld __cnfn bitselect(ulong4 a, ulong4 b, ulong4 c);
-long8 __ovld __cnfn bitselect(long8 a, long8 b, long8 c);
-ulong8 __ovld __cnfn bitselect(ulong8 a, ulong8 b, ulong8 c);
-long16 __ovld __cnfn bitselect(long16 a, long16 b, long16 c);
-ulong16 __ovld __cnfn bitselect(ulong16 a, ulong16 b, ulong16 c);
-float __ovld __cnfn bitselect(float a, float b, float c);
-float2 __ovld __cnfn bitselect(float2 a, float2 b, float2 c);
-float3 __ovld __cnfn bitselect(float3 a, float3 b, float3 c);
-float4 __ovld __cnfn bitselect(float4 a, float4 b, float4 c);
-float8 __ovld __cnfn bitselect(float8 a, float8 b, float8 c);
-float16 __ovld __cnfn bitselect(float16 a, float16 b, float16 c);
-#ifdef cl_khr_fp64
-double __ovld __cnfn bitselect(double a, double b, double c);
-double2 __ovld __cnfn bitselect(double2 a, double2 b, double2 c);
-double3 __ovld __cnfn bitselect(double3 a, double3 b, double3 c);
-double4 __ovld __cnfn bitselect(double4 a, double4 b, double4 c);
-double8 __ovld __cnfn bitselect(double8 a, double8 b, double8 c);
-double16 __ovld __cnfn bitselect(double16 a, double16 b, double16 c);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn bitselect(half a, half b, half c);
-half2 __ovld __cnfn bitselect(half2 a, half2 b, half2 c);
-half3 __ovld __cnfn bitselect(half3 a, half3 b, half3 c);
-half4 __ovld __cnfn bitselect(half4 a, half4 b, half4 c);
-half8 __ovld __cnfn bitselect(half8 a, half8 b, half8 c);
-half16 __ovld __cnfn bitselect(half16 a, half16 b, half16 c);
-#endif //cl_khr_fp16
-
-/**
- * For each component of a vector type,
- * result[i] = if MSB of c[i] is set ? b[i] : a[i].
- * For a scalar type, result = c ? b : a.
- * b and a must have the same type.
- * c must have the same number of elements and bits as a.
- */
-char __ovld __cnfn select(char a, char b, char c);
-uchar __ovld __cnfn select(uchar a, uchar b, char c);
-char2 __ovld __cnfn select(char2 a, char2 b, char2 c);
-uchar2 __ovld __cnfn select(uchar2 a, uchar2 b, char2 c);
-char3 __ovld __cnfn select(char3 a, char3 b, char3 c);
-uchar3 __ovld __cnfn select(uchar3 a, uchar3 b, char3 c);
-char4 __ovld __cnfn select(char4 a, char4 b, char4 c);
-uchar4 __ovld __cnfn select(uchar4 a, uchar4 b, char4 c);
-char8 __ovld __cnfn select(char8 a, char8 b, char8 c);
-uchar8 __ovld __cnfn select(uchar8 a, uchar8 b, char8 c);
-char16 __ovld __cnfn select(char16 a, char16 b, char16 c);
-uchar16 __ovld __cnfn select(uchar16 a, uchar16 b, char16 c);
-
-short __ovld __cnfn select(short a, short b, short c);
-ushort __ovld __cnfn select(ushort a, ushort b, short c);
-short2 __ovld __cnfn select(short2 a, short2 b, short2 c);
-ushort2 __ovld __cnfn select(ushort2 a, ushort2 b, short2 c);
-short3 __ovld __cnfn select(short3 a, short3 b, short3 c);
-ushort3 __ovld __cnfn select(ushort3 a, ushort3 b, short3 c);
-short4 __ovld __cnfn select(short4 a, short4 b, short4 c);
-ushort4 __ovld __cnfn select(ushort4 a, ushort4 b, short4 c);
-short8 __ovld __cnfn select(short8 a, short8 b, short8 c);
-ushort8 __ovld __cnfn select(ushort8 a, ushort8 b, short8 c);
-short16 __ovld __cnfn select(short16 a, short16 b, short16 c);
-ushort16 __ovld __cnfn select(ushort16 a, ushort16 b, short16 c);
-
-int __ovld __cnfn select(int a, int b, int c);
-uint __ovld __cnfn select(uint a, uint b, int c);
-int2 __ovld __cnfn select(int2 a, int2 b, int2 c);
-uint2 __ovld __cnfn select(uint2 a, uint2 b, int2 c);
-int3 __ovld __cnfn select(int3 a, int3 b, int3 c);
-uint3 __ovld __cnfn select(uint3 a, uint3 b, int3 c);
-int4 __ovld __cnfn select(int4 a, int4 b, int4 c);
-uint4 __ovld __cnfn select(uint4 a, uint4 b, int4 c);
-int8 __ovld __cnfn select(int8 a, int8 b, int8 c);
-uint8 __ovld __cnfn select(uint8 a, uint8 b, int8 c);
-int16 __ovld __cnfn select(int16 a, int16 b, int16 c);
-uint16 __ovld __cnfn select(uint16 a, uint16 b, int16 c);
-float __ovld __cnfn select(float a, float b, int c);
-float2 __ovld __cnfn select(float2 a, float2 b, int2 c);
-float3 __ovld __cnfn select(float3 a, float3 b, int3 c);
-float4 __ovld __cnfn select(float4 a, float4 b, int4 c);
-float8 __ovld __cnfn select(float8 a, float8 b, int8 c);
-float16 __ovld __cnfn select(float16 a, float16 b, int16 c);
-
-long __ovld __cnfn select(long a, long b, long c);
-ulong __ovld __cnfn select(ulong a, ulong b, long c);
-long2 __ovld __cnfn select(long2 a, long2 b, long2 c);
-ulong2 __ovld __cnfn select(ulong2 a, ulong2 b, long2 c);
-long3 __ovld __cnfn select(long3 a, long3 b, long3 c);
-ulong3 __ovld __cnfn select(ulong3 a, ulong3 b, long3 c);
-long4 __ovld __cnfn select(long4 a, long4 b, long4 c);
-ulong4 __ovld __cnfn select(ulong4 a, ulong4 b, long4 c);
-long8 __ovld __cnfn select(long8 a, long8 b, long8 c);
-ulong8 __ovld __cnfn select(ulong8 a, ulong8 b, long8 c);
-long16 __ovld __cnfn select(long16 a, long16 b, long16 c);
-ulong16 __ovld __cnfn select(ulong16 a, ulong16 b, long16 c);
-
-char __ovld __cnfn select(char a, char b, uchar c);
-uchar __ovld __cnfn select(uchar a, uchar b, uchar c);
-char2 __ovld __cnfn select(char2 a, char2 b, uchar2 c);
-uchar2 __ovld __cnfn select(uchar2 a, uchar2 b, uchar2 c);
-char3 __ovld __cnfn select(char3 a, char3 b, uchar3 c);
-uchar3 __ovld __cnfn select(uchar3 a, uchar3 b, uchar3 c);
-char4 __ovld __cnfn select(char4 a, char4 b, uchar4 c);
-uchar4 __ovld __cnfn select(uchar4 a, uchar4 b, uchar4 c);
-char8 __ovld __cnfn select(char8 a, char8 b, uchar8 c);
-uchar8 __ovld __cnfn select(uchar8 a, uchar8 b, uchar8 c);
-char16 __ovld __cnfn select(char16 a, char16 b, uchar16 c);
-uchar16 __ovld __cnfn select(uchar16 a, uchar16 b, uchar16 c);
-
-short __ovld __cnfn select(short a, short b, ushort c);
-ushort __ovld __cnfn select(ushort a, ushort b, ushort c);
-short2 __ovld __cnfn select(short2 a, short2 b, ushort2 c);
-ushort2 __ovld __cnfn select(ushort2 a, ushort2 b, ushort2 c);
-short3 __ovld __cnfn select(short3 a, short3 b, ushort3 c);
-ushort3 __ovld __cnfn select(ushort3 a, ushort3 b, ushort3 c);
-short4 __ovld __cnfn select(short4 a, short4 b, ushort4 c);
-ushort4 __ovld __cnfn select(ushort4 a, ushort4 b, ushort4 c);
-short8 __ovld __cnfn select(short8 a, short8 b, ushort8 c);
-ushort8 __ovld __cnfn select(ushort8 a, ushort8 b, ushort8 c);
-short16 __ovld __cnfn select(short16 a, short16 b, ushort16 c);
-ushort16 __ovld __cnfn select(ushort16 a, ushort16 b, ushort16 c);
-
-int __ovld __cnfn select(int a, int b, uint c);
-uint __ovld __cnfn select(uint a, uint b, uint c);
-int2 __ovld __cnfn select(int2 a, int2 b, uint2 c);
-uint2 __ovld __cnfn select(uint2 a, uint2 b, uint2 c);
-int3 __ovld __cnfn select(int3 a, int3 b, uint3 c);
-uint3 __ovld __cnfn select(uint3 a, uint3 b, uint3 c);
-int4 __ovld __cnfn select(int4 a, int4 b, uint4 c);
-uint4 __ovld __cnfn select(uint4 a, uint4 b, uint4 c);
-int8 __ovld __cnfn select(int8 a, int8 b, uint8 c);
-uint8 __ovld __cnfn select(uint8 a, uint8 b, uint8 c);
-int16 __ovld __cnfn select(int16 a, int16 b, uint16 c);
-uint16 __ovld __cnfn select(uint16 a, uint16 b, uint16 c);
-float __ovld __cnfn select(float a, float b, uint c);
-float2 __ovld __cnfn select(float2 a, float2 b, uint2 c);
-float3 __ovld __cnfn select(float3 a, float3 b, uint3 c);
-float4 __ovld __cnfn select(float4 a, float4 b, uint4 c);
-float8 __ovld __cnfn select(float8 a, float8 b, uint8 c);
-float16 __ovld __cnfn select(float16 a, float16 b, uint16 c);
-
-long __ovld __cnfn select(long a, long b, ulong c);
-ulong __ovld __cnfn select(ulong a, ulong b, ulong c);
-long2 __ovld __cnfn select(long2 a, long2 b, ulong2 c);
-ulong2 __ovld __cnfn select(ulong2 a, ulong2 b, ulong2 c);
-long3 __ovld __cnfn select(long3 a, long3 b, ulong3 c);
-ulong3 __ovld __cnfn select(ulong3 a, ulong3 b, ulong3 c);
-long4 __ovld __cnfn select(long4 a, long4 b, ulong4 c);
-ulong4 __ovld __cnfn select(ulong4 a, ulong4 b, ulong4 c);
-long8 __ovld __cnfn select(long8 a, long8 b, ulong8 c);
-ulong8 __ovld __cnfn select(ulong8 a, ulong8 b, ulong8 c);
-long16 __ovld __cnfn select(long16 a, long16 b, ulong16 c);
-ulong16 __ovld __cnfn select(ulong16 a, ulong16 b, ulong16 c);
-
-#ifdef cl_khr_fp64
-double __ovld __cnfn select(double a, double b, long c);
-double2 __ovld __cnfn select(double2 a, double2 b, long2 c);
-double3 __ovld __cnfn select(double3 a, double3 b, long3 c);
-double4 __ovld __cnfn select(double4 a, double4 b, long4 c);
-double8 __ovld __cnfn select(double8 a, double8 b, long8 c);
-double16 __ovld __cnfn select(double16 a, double16 b, long16 c);
-double __ovld __cnfn select(double a, double b, ulong c);
-double2 __ovld __cnfn select(double2 a, double2 b, ulong2 c);
-double3 __ovld __cnfn select(double3 a, double3 b, ulong3 c);
-double4 __ovld __cnfn select(double4 a, double4 b, ulong4 c);
-double8 __ovld __cnfn select(double8 a, double8 b, ulong8 c);
-double16 __ovld __cnfn select(double16 a, double16 b, ulong16 c);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-half __ovld __cnfn select(half a, half b, short c);
-half2 __ovld __cnfn select(half2 a, half2 b, short2 c);
-half3 __ovld __cnfn select(half3 a, half3 b, short3 c);
-half4 __ovld __cnfn select(half4 a, half4 b, short4 c);
-half8 __ovld __cnfn select(half8 a, half8 b, short8 c);
-half16 __ovld __cnfn select(half16 a, half16 b, short16 c);
-half __ovld __cnfn select(half a, half b, ushort c);
-half2 __ovld __cnfn select(half2 a, half2 b, ushort2 c);
-half3 __ovld __cnfn select(half3 a, half3 b, ushort3 c);
-half4 __ovld __cnfn select(half4 a, half4 b, ushort4 c);
-half8 __ovld __cnfn select(half8 a, half8 b, ushort8 c);
-half16 __ovld __cnfn select(half16 a, half16 b, ushort16 c);
-#endif //cl_khr_fp16
-
-// OpenCL v1.1 s6.11.7, v1.2 s6.12.7, v2.0 s6.13.7 - Vector Data Load and Store Functions
-// OpenCL extensions v1.1 s9.6.6, v1.2 s9.5.6, v2.0 s9.4.6 - Vector Data Load and Store Functions for Half Type
-/**
- * Use generic type gentype to indicate the built-in data types
- * char, uchar, short, ushort, int, uint, long, ulong, float,
- * double or half.
- *
- * vloadn return sizeof (gentypen) bytes of data read from address (p + (offset * n)).
- *
- * vstoren write sizeof (gentypen) bytes given by data to address (p + (offset * n)).
- *
- * The address computed as (p + (offset * n)) must be
- * 8-bit aligned if gentype is char, uchar;
- * 16-bit aligned if gentype is short, ushort, half;
- * 32-bit aligned if gentype is int, uint, float;
- * 64-bit aligned if gentype is long, ulong, double.
- */
-
-char2 __ovld vload2(size_t offset, const __constant char *p);
-uchar2 __ovld vload2(size_t offset, const __constant uchar *p);
-short2 __ovld vload2(size_t offset, const __constant short *p);
-ushort2 __ovld vload2(size_t offset, const __constant ushort *p);
-int2 __ovld vload2(size_t offset, const __constant int *p);
-uint2 __ovld vload2(size_t offset, const __constant uint *p);
-long2 __ovld vload2(size_t offset, const __constant long *p);
-ulong2 __ovld vload2(size_t offset, const __constant ulong *p);
-float2 __ovld vload2(size_t offset, const __constant float *p);
-char3 __ovld vload3(size_t offset, const __constant char *p);
-uchar3 __ovld vload3(size_t offset, const __constant uchar *p);
-short3 __ovld vload3(size_t offset, const __constant short *p);
-ushort3 __ovld vload3(size_t offset, const __constant ushort *p);
-int3 __ovld vload3(size_t offset, const __constant int *p);
-uint3 __ovld vload3(size_t offset, const __constant uint *p);
-long3 __ovld vload3(size_t offset, const __constant long *p);
-ulong3 __ovld vload3(size_t offset, const __constant ulong *p);
-float3 __ovld vload3(size_t offset, const __constant float *p);
-char4 __ovld vload4(size_t offset, const __constant char *p);
-uchar4 __ovld vload4(size_t offset, const __constant uchar *p);
-short4 __ovld vload4(size_t offset, const __constant short *p);
-ushort4 __ovld vload4(size_t offset, const __constant ushort *p);
-int4 __ovld vload4(size_t offset, const __constant int *p);
-uint4 __ovld vload4(size_t offset, const __constant uint *p);
-long4 __ovld vload4(size_t offset, const __constant long *p);
-ulong4 __ovld vload4(size_t offset, const __constant ulong *p);
-float4 __ovld vload4(size_t offset, const __constant float *p);
-char8 __ovld vload8(size_t offset, const __constant char *p);
-uchar8 __ovld vload8(size_t offset, const __constant uchar *p);
-short8 __ovld vload8(size_t offset, const __constant short *p);
-ushort8 __ovld vload8(size_t offset, const __constant ushort *p);
-int8 __ovld vload8(size_t offset, const __constant int *p);
-uint8 __ovld vload8(size_t offset, const __constant uint *p);
-long8 __ovld vload8(size_t offset, const __constant long *p);
-ulong8 __ovld vload8(size_t offset, const __constant ulong *p);
-float8 __ovld vload8(size_t offset, const __constant float *p);
-char16 __ovld vload16(size_t offset, const __constant char *p);
-uchar16 __ovld vload16(size_t offset, const __constant uchar *p);
-short16 __ovld vload16(size_t offset, const __constant short *p);
-ushort16 __ovld vload16(size_t offset, const __constant ushort *p);
-int16 __ovld vload16(size_t offset, const __constant int *p);
-uint16 __ovld vload16(size_t offset, const __constant uint *p);
-long16 __ovld vload16(size_t offset, const __constant long *p);
-ulong16 __ovld vload16(size_t offset, const __constant ulong *p);
-float16 __ovld vload16(size_t offset, const __constant float *p);
-#ifdef cl_khr_fp64
-double2 __ovld vload2(size_t offset, const __constant double *p);
-double3 __ovld vload3(size_t offset, const __constant double *p);
-double4 __ovld vload4(size_t offset, const __constant double *p);
-double8 __ovld vload8(size_t offset, const __constant double *p);
-double16 __ovld vload16(size_t offset, const __constant double *p);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half __ovld vload(size_t offset, const __constant half *p);
-half2 __ovld vload2(size_t offset, const __constant half *p);
-half3 __ovld vload3(size_t offset, const __constant half *p);
-half4 __ovld vload4(size_t offset, const __constant half *p);
-half8 __ovld vload8(size_t offset, const __constant half *p);
-half16 __ovld vload16(size_t offset, const __constant half *p);
-#endif //cl_khr_fp16
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-char2 __ovld vload2(size_t offset, const char *p);
-uchar2 __ovld vload2(size_t offset, const uchar *p);
-short2 __ovld vload2(size_t offset, const short *p);
-ushort2 __ovld vload2(size_t offset, const ushort *p);
-int2 __ovld vload2(size_t offset, const int *p);
-uint2 __ovld vload2(size_t offset, const uint *p);
-long2 __ovld vload2(size_t offset, const long *p);
-ulong2 __ovld vload2(size_t offset, const ulong *p);
-float2 __ovld vload2(size_t offset, const float *p);
-char3 __ovld vload3(size_t offset, const char *p);
-uchar3 __ovld vload3(size_t offset, const uchar *p);
-short3 __ovld vload3(size_t offset, const short *p);
-ushort3 __ovld vload3(size_t offset, const ushort *p);
-int3 __ovld vload3(size_t offset, const int *p);
-uint3 __ovld vload3(size_t offset, const uint *p);
-long3 __ovld vload3(size_t offset, const long *p);
-ulong3 __ovld vload3(size_t offset, const ulong *p);
-float3 __ovld vload3(size_t offset, const float *p);
-char4 __ovld vload4(size_t offset, const char *p);
-uchar4 __ovld vload4(size_t offset, const uchar *p);
-short4 __ovld vload4(size_t offset, const short *p);
-ushort4 __ovld vload4(size_t offset, const ushort *p);
-int4 __ovld vload4(size_t offset, const int *p);
-uint4 __ovld vload4(size_t offset, const uint *p);
-long4 __ovld vload4(size_t offset, const long *p);
-ulong4 __ovld vload4(size_t offset, const ulong *p);
-float4 __ovld vload4(size_t offset, const float *p);
-char8 __ovld vload8(size_t offset, const char *p);
-uchar8 __ovld vload8(size_t offset, const uchar *p);
-short8 __ovld vload8(size_t offset, const short *p);
-ushort8 __ovld vload8(size_t offset, const ushort *p);
-int8 __ovld vload8(size_t offset, const int *p);
-uint8 __ovld vload8(size_t offset, const uint *p);
-long8 __ovld vload8(size_t offset, const long *p);
-ulong8 __ovld vload8(size_t offset, const ulong *p);
-float8 __ovld vload8(size_t offset, const float *p);
-char16 __ovld vload16(size_t offset, const char *p);
-uchar16 __ovld vload16(size_t offset, const uchar *p);
-short16 __ovld vload16(size_t offset, const short *p);
-ushort16 __ovld vload16(size_t offset, const ushort *p);
-int16 __ovld vload16(size_t offset, const int *p);
-uint16 __ovld vload16(size_t offset, const uint *p);
-long16 __ovld vload16(size_t offset, const long *p);
-ulong16 __ovld vload16(size_t offset, const ulong *p);
-float16 __ovld vload16(size_t offset, const float *p);
-
-#ifdef cl_khr_fp64
-double2 __ovld vload2(size_t offset, const double *p);
-double3 __ovld vload3(size_t offset, const double *p);
-double4 __ovld vload4(size_t offset, const double *p);
-double8 __ovld vload8(size_t offset, const double *p);
-double16 __ovld vload16(size_t offset, const double *p);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half __ovld vload(size_t offset, const half *p);
-half2 __ovld vload2(size_t offset, const half *p);
-half3 __ovld vload3(size_t offset, const half *p);
-half4 __ovld vload4(size_t offset, const half *p);
-half8 __ovld vload8(size_t offset, const half *p);
-half16 __ovld vload16(size_t offset, const half *p);
-#endif //cl_khr_fp16
-#else
-char2 __ovld vload2(size_t offset, const __global char *p);
-uchar2 __ovld vload2(size_t offset, const __global uchar *p);
-short2 __ovld vload2(size_t offset, const __global short *p);
-ushort2 __ovld vload2(size_t offset, const __global ushort *p);
-int2 __ovld vload2(size_t offset, const __global int *p);
-uint2 __ovld vload2(size_t offset, const __global uint *p);
-long2 __ovld vload2(size_t offset, const __global long *p);
-ulong2 __ovld vload2(size_t offset, const __global ulong *p);
-float2 __ovld vload2(size_t offset, const __global float *p);
-char3 __ovld vload3(size_t offset, const __global char *p);
-uchar3 __ovld vload3(size_t offset, const __global uchar *p);
-short3 __ovld vload3(size_t offset, const __global short *p);
-ushort3 __ovld vload3(size_t offset, const __global ushort *p);
-int3 __ovld vload3(size_t offset, const __global int *p);
-uint3 __ovld vload3(size_t offset, const __global uint *p);
-long3 __ovld vload3(size_t offset, const __global long *p);
-ulong3 __ovld vload3(size_t offset, const __global ulong *p);
-float3 __ovld vload3(size_t offset, const __global float *p);
-char4 __ovld vload4(size_t offset, const __global char *p);
-uchar4 __ovld vload4(size_t offset, const __global uchar *p);
-short4 __ovld vload4(size_t offset, const __global short *p);
-ushort4 __ovld vload4(size_t offset, const __global ushort *p);
-int4 __ovld vload4(size_t offset, const __global int *p);
-uint4 __ovld vload4(size_t offset, const __global uint *p);
-long4 __ovld vload4(size_t offset, const __global long *p);
-ulong4 __ovld vload4(size_t offset, const __global ulong *p);
-float4 __ovld vload4(size_t offset, const __global float *p);
-char8 __ovld vload8(size_t offset, const __global char *p);
-uchar8 __ovld vload8(size_t offset, const __global uchar *p);
-short8 __ovld vload8(size_t offset, const __global short *p);
-ushort8 __ovld vload8(size_t offset, const __global ushort *p);
-int8 __ovld vload8(size_t offset, const __global int *p);
-uint8 __ovld vload8(size_t offset, const __global uint *p);
-long8 __ovld vload8(size_t offset, const __global long *p);
-ulong8 __ovld vload8(size_t offset, const __global ulong *p);
-float8 __ovld vload8(size_t offset, const __global float *p);
-char16 __ovld vload16(size_t offset, const __global char *p);
-uchar16 __ovld vload16(size_t offset, const __global uchar *p);
-short16 __ovld vload16(size_t offset, const __global short *p);
-ushort16 __ovld vload16(size_t offset, const __global ushort *p);
-int16 __ovld vload16(size_t offset, const __global int *p);
-uint16 __ovld vload16(size_t offset, const __global uint *p);
-long16 __ovld vload16(size_t offset, const __global long *p);
-ulong16 __ovld vload16(size_t offset, const __global ulong *p);
-float16 __ovld vload16(size_t offset, const __global float *p);
-char2 __ovld vload2(size_t offset, const __local char *p);
-uchar2 __ovld vload2(size_t offset, const __local uchar *p);
-short2 __ovld vload2(size_t offset, const __local short *p);
-ushort2 __ovld vload2(size_t offset, const __local ushort *p);
-int2 __ovld vload2(size_t offset, const __local int *p);
-uint2 __ovld vload2(size_t offset, const __local uint *p);
-long2 __ovld vload2(size_t offset, const __local long *p);
-ulong2 __ovld vload2(size_t offset, const __local ulong *p);
-float2 __ovld vload2(size_t offset, const __local float *p);
-char3 __ovld vload3(size_t offset, const __local char *p);
-uchar3 __ovld vload3(size_t offset, const __local uchar *p);
-short3 __ovld vload3(size_t offset, const __local short *p);
-ushort3 __ovld vload3(size_t offset, const __local ushort *p);
-int3 __ovld vload3(size_t offset, const __local int *p);
-uint3 __ovld vload3(size_t offset, const __local uint *p);
-long3 __ovld vload3(size_t offset, const __local long *p);
-ulong3 __ovld vload3(size_t offset, const __local ulong *p);
-float3 __ovld vload3(size_t offset, const __local float *p);
-char4 __ovld vload4(size_t offset, const __local char *p);
-uchar4 __ovld vload4(size_t offset, const __local uchar *p);
-short4 __ovld vload4(size_t offset, const __local short *p);
-ushort4 __ovld vload4(size_t offset, const __local ushort *p);
-int4 __ovld vload4(size_t offset, const __local int *p);
-uint4 __ovld vload4(size_t offset, const __local uint *p);
-long4 __ovld vload4(size_t offset, const __local long *p);
-ulong4 __ovld vload4(size_t offset, const __local ulong *p);
-float4 __ovld vload4(size_t offset, const __local float *p);
-char8 __ovld vload8(size_t offset, const __local char *p);
-uchar8 __ovld vload8(size_t offset, const __local uchar *p);
-short8 __ovld vload8(size_t offset, const __local short *p);
-ushort8 __ovld vload8(size_t offset, const __local ushort *p);
-int8 __ovld vload8(size_t offset, const __local int *p);
-uint8 __ovld vload8(size_t offset, const __local uint *p);
-long8 __ovld vload8(size_t offset, const __local long *p);
-ulong8 __ovld vload8(size_t offset, const __local ulong *p);
-float8 __ovld vload8(size_t offset, const __local float *p);
-char16 __ovld vload16(size_t offset, const __local char *p);
-uchar16 __ovld vload16(size_t offset, const __local uchar *p);
-short16 __ovld vload16(size_t offset, const __local short *p);
-ushort16 __ovld vload16(size_t offset, const __local ushort *p);
-int16 __ovld vload16(size_t offset, const __local int *p);
-uint16 __ovld vload16(size_t offset, const __local uint *p);
-long16 __ovld vload16(size_t offset, const __local long *p);
-ulong16 __ovld vload16(size_t offset, const __local ulong *p);
-float16 __ovld vload16(size_t offset, const __local float *p);
-char2 __ovld vload2(size_t offset, const __private char *p);
-uchar2 __ovld vload2(size_t offset, const __private uchar *p);
-short2 __ovld vload2(size_t offset, const __private short *p);
-ushort2 __ovld vload2(size_t offset, const __private ushort *p);
-int2 __ovld vload2(size_t offset, const __private int *p);
-uint2 __ovld vload2(size_t offset, const __private uint *p);
-long2 __ovld vload2(size_t offset, const __private long *p);
-ulong2 __ovld vload2(size_t offset, const __private ulong *p);
-float2 __ovld vload2(size_t offset, const __private float *p);
-char3 __ovld vload3(size_t offset, const __private char *p);
-uchar3 __ovld vload3(size_t offset, const __private uchar *p);
-short3 __ovld vload3(size_t offset, const __private short *p);
-ushort3 __ovld vload3(size_t offset, const __private ushort *p);
-int3 __ovld vload3(size_t offset, const __private int *p);
-uint3 __ovld vload3(size_t offset, const __private uint *p);
-long3 __ovld vload3(size_t offset, const __private long *p);
-ulong3 __ovld vload3(size_t offset, const __private ulong *p);
-float3 __ovld vload3(size_t offset, const __private float *p);
-char4 __ovld vload4(size_t offset, const __private char *p);
-uchar4 __ovld vload4(size_t offset, const __private uchar *p);
-short4 __ovld vload4(size_t offset, const __private short *p);
-ushort4 __ovld vload4(size_t offset, const __private ushort *p);
-int4 __ovld vload4(size_t offset, const __private int *p);
-uint4 __ovld vload4(size_t offset, const __private uint *p);
-long4 __ovld vload4(size_t offset, const __private long *p);
-ulong4 __ovld vload4(size_t offset, const __private ulong *p);
-float4 __ovld vload4(size_t offset, const __private float *p);
-char8 __ovld vload8(size_t offset, const __private char *p);
-uchar8 __ovld vload8(size_t offset, const __private uchar *p);
-short8 __ovld vload8(size_t offset, const __private short *p);
-ushort8 __ovld vload8(size_t offset, const __private ushort *p);
-int8 __ovld vload8(size_t offset, const __private int *p);
-uint8 __ovld vload8(size_t offset, const __private uint *p);
-long8 __ovld vload8(size_t offset, const __private long *p);
-ulong8 __ovld vload8(size_t offset, const __private ulong *p);
-float8 __ovld vload8(size_t offset, const __private float *p);
-char16 __ovld vload16(size_t offset, const __private char *p);
-uchar16 __ovld vload16(size_t offset, const __private uchar *p);
-short16 __ovld vload16(size_t offset, const __private short *p);
-ushort16 __ovld vload16(size_t offset, const __private ushort *p);
-int16 __ovld vload16(size_t offset, const __private int *p);
-uint16 __ovld vload16(size_t offset, const __private uint *p);
-long16 __ovld vload16(size_t offset, const __private long *p);
-ulong16 __ovld vload16(size_t offset, const __private ulong *p);
-float16 __ovld vload16(size_t offset, const __private float *p);
-
-#ifdef cl_khr_fp64
-double2 __ovld vload2(size_t offset, const __global double *p);
-double3 __ovld vload3(size_t offset, const __global double *p);
-double4 __ovld vload4(size_t offset, const __global double *p);
-double8 __ovld vload8(size_t offset, const __global double *p);
-double16 __ovld vload16(size_t offset, const __global double *p);
-double2 __ovld vload2(size_t offset, const __local double *p);
-double3 __ovld vload3(size_t offset, const __local double *p);
-double4 __ovld vload4(size_t offset, const __local double *p);
-double8 __ovld vload8(size_t offset, const __local double *p);
-double16 __ovld vload16(size_t offset, const __local double *p);
-double2 __ovld vload2(size_t offset, const __private double *p);
-double3 __ovld vload3(size_t offset, const __private double *p);
-double4 __ovld vload4(size_t offset, const __private double *p);
-double8 __ovld vload8(size_t offset, const __private double *p);
-double16 __ovld vload16(size_t offset, const __private double *p);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half __ovld vload(size_t offset, const __global half *p);
-half2 __ovld vload2(size_t offset, const __global half *p);
-half3 __ovld vload3(size_t offset, const __global half *p);
-half4 __ovld vload4(size_t offset, const __global half *p);
-half8 __ovld vload8(size_t offset, const __global half *p);
-half16 __ovld vload16(size_t offset, const __global half *p);
-half __ovld vload(size_t offset, const __local half *p);
-half2 __ovld vload2(size_t offset, const __local half *p);
-half3 __ovld vload3(size_t offset, const __local half *p);
-half4 __ovld vload4(size_t offset, const __local half *p);
-half8 __ovld vload8(size_t offset, const __local half *p);
-half16 __ovld vload16(size_t offset, const __local half *p);
-half __ovld vload(size_t offset, const __private half *p);
-half2 __ovld vload2(size_t offset, const __private half *p);
-half3 __ovld vload3(size_t offset, const __private half *p);
-half4 __ovld vload4(size_t offset, const __private half *p);
-half8 __ovld vload8(size_t offset, const __private half *p);
-half16 __ovld vload16(size_t offset, const __private half *p);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld vstore2(char2 data, size_t offset, char *p);
-void __ovld vstore2(uchar2 data, size_t offset, uchar *p);
-void __ovld vstore2(short2 data, size_t offset, short *p);
-void __ovld vstore2(ushort2 data, size_t offset, ushort *p);
-void __ovld vstore2(int2 data, size_t offset, int *p);
-void __ovld vstore2(uint2 data, size_t offset, uint *p);
-void __ovld vstore2(long2 data, size_t offset, long *p);
-void __ovld vstore2(ulong2 data, size_t offset, ulong *p);
-void __ovld vstore2(float2 data, size_t offset, float *p);
-void __ovld vstore3(char3 data, size_t offset, char *p);
-void __ovld vstore3(uchar3 data, size_t offset, uchar *p);
-void __ovld vstore3(short3 data, size_t offset, short *p);
-void __ovld vstore3(ushort3 data, size_t offset, ushort *p);
-void __ovld vstore3(int3 data, size_t offset, int *p);
-void __ovld vstore3(uint3 data, size_t offset, uint *p);
-void __ovld vstore3(long3 data, size_t offset, long *p);
-void __ovld vstore3(ulong3 data, size_t offset, ulong *p);
-void __ovld vstore3(float3 data, size_t offset, float *p);
-void __ovld vstore4(char4 data, size_t offset, char *p);
-void __ovld vstore4(uchar4 data, size_t offset, uchar *p);
-void __ovld vstore4(short4 data, size_t offset, short *p);
-void __ovld vstore4(ushort4 data, size_t offset, ushort *p);
-void __ovld vstore4(int4 data, size_t offset, int *p);
-void __ovld vstore4(uint4 data, size_t offset, uint *p);
-void __ovld vstore4(long4 data, size_t offset, long *p);
-void __ovld vstore4(ulong4 data, size_t offset, ulong *p);
-void __ovld vstore4(float4 data, size_t offset, float *p);
-void __ovld vstore8(char8 data, size_t offset, char *p);
-void __ovld vstore8(uchar8 data, size_t offset, uchar *p);
-void __ovld vstore8(short8 data, size_t offset, short *p);
-void __ovld vstore8(ushort8 data, size_t offset, ushort *p);
-void __ovld vstore8(int8 data, size_t offset, int *p);
-void __ovld vstore8(uint8 data, size_t offset, uint *p);
-void __ovld vstore8(long8 data, size_t offset, long *p);
-void __ovld vstore8(ulong8 data, size_t offset, ulong *p);
-void __ovld vstore8(float8 data, size_t offset, float *p);
-void __ovld vstore16(char16 data, size_t offset, char *p);
-void __ovld vstore16(uchar16 data, size_t offset, uchar *p);
-void __ovld vstore16(short16 data, size_t offset, short *p);
-void __ovld vstore16(ushort16 data, size_t offset, ushort *p);
-void __ovld vstore16(int16 data, size_t offset, int *p);
-void __ovld vstore16(uint16 data, size_t offset, uint *p);
-void __ovld vstore16(long16 data, size_t offset, long *p);
-void __ovld vstore16(ulong16 data, size_t offset, ulong *p);
-void __ovld vstore16(float16 data, size_t offset, float *p);
-#ifdef cl_khr_fp64
-void __ovld vstore2(double2 data, size_t offset, double *p);
-void __ovld vstore3(double3 data, size_t offset, double *p);
-void __ovld vstore4(double4 data, size_t offset, double *p);
-void __ovld vstore8(double8 data, size_t offset, double *p);
-void __ovld vstore16(double16 data, size_t offset, double *p);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-void __ovld vstore(half data, size_t offset, half *p);
-void __ovld vstore2(half2 data, size_t offset, half *p);
-void __ovld vstore3(half3 data, size_t offset, half *p);
-void __ovld vstore4(half4 data, size_t offset, half *p);
-void __ovld vstore8(half8 data, size_t offset, half *p);
-void __ovld vstore16(half16 data, size_t offset, half *p);
-#endif //cl_khr_fp16
-#else
-void __ovld vstore2(char2 data, size_t offset, __global char *p);
-void __ovld vstore2(uchar2 data, size_t offset, __global uchar *p);
-void __ovld vstore2(short2 data, size_t offset, __global short *p);
-void __ovld vstore2(ushort2 data, size_t offset, __global ushort *p);
-void __ovld vstore2(int2 data, size_t offset, __global int *p);
-void __ovld vstore2(uint2 data, size_t offset, __global uint *p);
-void __ovld vstore2(long2 data, size_t offset, __global long *p);
-void __ovld vstore2(ulong2 data, size_t offset, __global ulong *p);
-void __ovld vstore2(float2 data, size_t offset, __global float *p);
-void __ovld vstore3(char3 data, size_t offset, __global char *p);
-void __ovld vstore3(uchar3 data, size_t offset, __global uchar *p);
-void __ovld vstore3(short3 data, size_t offset, __global short *p);
-void __ovld vstore3(ushort3 data, size_t offset, __global ushort *p);
-void __ovld vstore3(int3 data, size_t offset, __global int *p);
-void __ovld vstore3(uint3 data, size_t offset, __global uint *p);
-void __ovld vstore3(long3 data, size_t offset, __global long *p);
-void __ovld vstore3(ulong3 data, size_t offset, __global ulong *p);
-void __ovld vstore3(float3 data, size_t offset, __global float *p);
-void __ovld vstore4(char4 data, size_t offset, __global char *p);
-void __ovld vstore4(uchar4 data, size_t offset, __global uchar *p);
-void __ovld vstore4(short4 data, size_t offset, __global short *p);
-void __ovld vstore4(ushort4 data, size_t offset, __global ushort *p);
-void __ovld vstore4(int4 data, size_t offset, __global int *p);
-void __ovld vstore4(uint4 data, size_t offset, __global uint *p);
-void __ovld vstore4(long4 data, size_t offset, __global long *p);
-void __ovld vstore4(ulong4 data, size_t offset, __global ulong *p);
-void __ovld vstore4(float4 data, size_t offset, __global float *p);
-void __ovld vstore8(char8 data, size_t offset, __global char *p);
-void __ovld vstore8(uchar8 data, size_t offset, __global uchar *p);
-void __ovld vstore8(short8 data, size_t offset, __global short *p);
-void __ovld vstore8(ushort8 data, size_t offset, __global ushort *p);
-void __ovld vstore8(int8 data, size_t offset, __global int *p);
-void __ovld vstore8(uint8 data, size_t offset, __global uint *p);
-void __ovld vstore8(long8 data, size_t offset, __global long *p);
-void __ovld vstore8(ulong8 data, size_t offset, __global ulong *p);
-void __ovld vstore8(float8 data, size_t offset, __global float *p);
-void __ovld vstore16(char16 data, size_t offset, __global char *p);
-void __ovld vstore16(uchar16 data, size_t offset, __global uchar *p);
-void __ovld vstore16(short16 data, size_t offset, __global short *p);
-void __ovld vstore16(ushort16 data, size_t offset, __global ushort *p);
-void __ovld vstore16(int16 data, size_t offset, __global int *p);
-void __ovld vstore16(uint16 data, size_t offset, __global uint *p);
-void __ovld vstore16(long16 data, size_t offset, __global long *p);
-void __ovld vstore16(ulong16 data, size_t offset, __global ulong *p);
-void __ovld vstore16(float16 data, size_t offset, __global float *p);
-void __ovld vstore2(char2 data, size_t offset, __local char *p);
-void __ovld vstore2(uchar2 data, size_t offset, __local uchar *p);
-void __ovld vstore2(short2 data, size_t offset, __local short *p);
-void __ovld vstore2(ushort2 data, size_t offset, __local ushort *p);
-void __ovld vstore2(int2 data, size_t offset, __local int *p);
-void __ovld vstore2(uint2 data, size_t offset, __local uint *p);
-void __ovld vstore2(long2 data, size_t offset, __local long *p);
-void __ovld vstore2(ulong2 data, size_t offset, __local ulong *p);
-void __ovld vstore2(float2 data, size_t offset, __local float *p);
-void __ovld vstore3(char3 data, size_t offset, __local char *p);
-void __ovld vstore3(uchar3 data, size_t offset, __local uchar *p);
-void __ovld vstore3(short3 data, size_t offset, __local short *p);
-void __ovld vstore3(ushort3 data, size_t offset, __local ushort *p);
-void __ovld vstore3(int3 data, size_t offset, __local int *p);
-void __ovld vstore3(uint3 data, size_t offset, __local uint *p);
-void __ovld vstore3(long3 data, size_t offset, __local long *p);
-void __ovld vstore3(ulong3 data, size_t offset, __local ulong *p);
-void __ovld vstore3(float3 data, size_t offset, __local float *p);
-void __ovld vstore4(char4 data, size_t offset, __local char *p);
-void __ovld vstore4(uchar4 data, size_t offset, __local uchar *p);
-void __ovld vstore4(short4 data, size_t offset, __local short *p);
-void __ovld vstore4(ushort4 data, size_t offset, __local ushort *p);
-void __ovld vstore4(int4 data, size_t offset, __local int *p);
-void __ovld vstore4(uint4 data, size_t offset, __local uint *p);
-void __ovld vstore4(long4 data, size_t offset, __local long *p);
-void __ovld vstore4(ulong4 data, size_t offset, __local ulong *p);
-void __ovld vstore4(float4 data, size_t offset, __local float *p);
-void __ovld vstore8(char8 data, size_t offset, __local char *p);
-void __ovld vstore8(uchar8 data, size_t offset, __local uchar *p);
-void __ovld vstore8(short8 data, size_t offset, __local short *p);
-void __ovld vstore8(ushort8 data, size_t offset, __local ushort *p);
-void __ovld vstore8(int8 data, size_t offset, __local int *p);
-void __ovld vstore8(uint8 data, size_t offset, __local uint *p);
-void __ovld vstore8(long8 data, size_t offset, __local long *p);
-void __ovld vstore8(ulong8 data, size_t offset, __local ulong *p);
-void __ovld vstore8(float8 data, size_t offset, __local float *p);
-void __ovld vstore16(char16 data, size_t offset, __local char *p);
-void __ovld vstore16(uchar16 data, size_t offset, __local uchar *p);
-void __ovld vstore16(short16 data, size_t offset, __local short *p);
-void __ovld vstore16(ushort16 data, size_t offset, __local ushort *p);
-void __ovld vstore16(int16 data, size_t offset, __local int *p);
-void __ovld vstore16(uint16 data, size_t offset, __local uint *p);
-void __ovld vstore16(long16 data, size_t offset, __local long *p);
-void __ovld vstore16(ulong16 data, size_t offset, __local ulong *p);
-void __ovld vstore16(float16 data, size_t offset, __local float *p);
-void __ovld vstore2(char2 data, size_t offset, __private char *p);
-void __ovld vstore2(uchar2 data, size_t offset, __private uchar *p);
-void __ovld vstore2(short2 data, size_t offset, __private short *p);
-void __ovld vstore2(ushort2 data, size_t offset, __private ushort *p);
-void __ovld vstore2(int2 data, size_t offset, __private int *p);
-void __ovld vstore2(uint2 data, size_t offset, __private uint *p);
-void __ovld vstore2(long2 data, size_t offset, __private long *p);
-void __ovld vstore2(ulong2 data, size_t offset, __private ulong *p);
-void __ovld vstore2(float2 data, size_t offset, __private float *p);
-void __ovld vstore3(char3 data, size_t offset, __private char *p);
-void __ovld vstore3(uchar3 data, size_t offset, __private uchar *p);
-void __ovld vstore3(short3 data, size_t offset, __private short *p);
-void __ovld vstore3(ushort3 data, size_t offset, __private ushort *p);
-void __ovld vstore3(int3 data, size_t offset, __private int *p);
-void __ovld vstore3(uint3 data, size_t offset, __private uint *p);
-void __ovld vstore3(long3 data, size_t offset, __private long *p);
-void __ovld vstore3(ulong3 data, size_t offset, __private ulong *p);
-void __ovld vstore3(float3 data, size_t offset, __private float *p);
-void __ovld vstore4(char4 data, size_t offset, __private char *p);
-void __ovld vstore4(uchar4 data, size_t offset, __private uchar *p);
-void __ovld vstore4(short4 data, size_t offset, __private short *p);
-void __ovld vstore4(ushort4 data, size_t offset, __private ushort *p);
-void __ovld vstore4(int4 data, size_t offset, __private int *p);
-void __ovld vstore4(uint4 data, size_t offset, __private uint *p);
-void __ovld vstore4(long4 data, size_t offset, __private long *p);
-void __ovld vstore4(ulong4 data, size_t offset, __private ulong *p);
-void __ovld vstore4(float4 data, size_t offset, __private float *p);
-void __ovld vstore8(char8 data, size_t offset, __private char *p);
-void __ovld vstore8(uchar8 data, size_t offset, __private uchar *p);
-void __ovld vstore8(short8 data, size_t offset, __private short *p);
-void __ovld vstore8(ushort8 data, size_t offset, __private ushort *p);
-void __ovld vstore8(int8 data, size_t offset, __private int *p);
-void __ovld vstore8(uint8 data, size_t offset, __private uint *p);
-void __ovld vstore8(long8 data, size_t offset, __private long *p);
-void __ovld vstore8(ulong8 data, size_t offset, __private ulong *p);
-void __ovld vstore8(float8 data, size_t offset, __private float *p);
-void __ovld vstore16(char16 data, size_t offset, __private char *p);
-void __ovld vstore16(uchar16 data, size_t offset, __private uchar *p);
-void __ovld vstore16(short16 data, size_t offset, __private short *p);
-void __ovld vstore16(ushort16 data, size_t offset, __private ushort *p);
-void __ovld vstore16(int16 data, size_t offset, __private int *p);
-void __ovld vstore16(uint16 data, size_t offset, __private uint *p);
-void __ovld vstore16(long16 data, size_t offset, __private long *p);
-void __ovld vstore16(ulong16 data, size_t offset, __private ulong *p);
-void __ovld vstore16(float16 data, size_t offset, __private float *p);
-#ifdef cl_khr_fp64
-void __ovld vstore2(double2 data, size_t offset, __global double *p);
-void __ovld vstore3(double3 data, size_t offset, __global double *p);
-void __ovld vstore4(double4 data, size_t offset, __global double *p);
-void __ovld vstore8(double8 data, size_t offset, __global double *p);
-void __ovld vstore16(double16 data, size_t offset, __global double *p);
-void __ovld vstore2(double2 data, size_t offset, __local double *p);
-void __ovld vstore3(double3 data, size_t offset, __local double *p);
-void __ovld vstore4(double4 data, size_t offset, __local double *p);
-void __ovld vstore8(double8 data, size_t offset, __local double *p);
-void __ovld vstore16(double16 data, size_t offset, __local double *p);
-void __ovld vstore2(double2 data, size_t offset, __private double *p);
-void __ovld vstore3(double3 data, size_t offset, __private double *p);
-void __ovld vstore4(double4 data, size_t offset, __private double *p);
-void __ovld vstore8(double8 data, size_t offset, __private double *p);
-void __ovld vstore16(double16 data, size_t offset, __private double *p);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-void __ovld vstore(half data, size_t offset, __global half *p);
-void __ovld vstore2(half2 data, size_t offset, __global half *p);
-void __ovld vstore3(half3 data, size_t offset, __global half *p);
-void __ovld vstore4(half4 data, size_t offset, __global half *p);
-void __ovld vstore8(half8 data, size_t offset, __global half *p);
-void __ovld vstore16(half16 data, size_t offset, __global half *p);
-void __ovld vstore(half data, size_t offset, __local half *p);
-void __ovld vstore2(half2 data, size_t offset, __local half *p);
-void __ovld vstore3(half3 data, size_t offset, __local half *p);
-void __ovld vstore4(half4 data, size_t offset, __local half *p);
-void __ovld vstore8(half8 data, size_t offset, __local half *p);
-void __ovld vstore16(half16 data, size_t offset, __local half *p);
-void __ovld vstore(half data, size_t offset, __private half *p);
-void __ovld vstore2(half2 data, size_t offset, __private half *p);
-void __ovld vstore3(half3 data, size_t offset, __private half *p);
-void __ovld vstore4(half4 data, size_t offset, __private half *p);
-void __ovld vstore8(half8 data, size_t offset, __private half *p);
-void __ovld vstore16(half16 data, size_t offset, __private half *p);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Read sizeof (half) bytes of data from address
- * (p + offset). The data read is interpreted as a
- * half value. The half value is converted to a
- * float value and the float value is returned.
- * The read address computed as (p + offset)
- * must be 16-bit aligned.
- */
-float __ovld vload_half(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld vload_half(size_t offset, const half *p);
-#else
-float __ovld vload_half(size_t offset, const __global half *p);
-float __ovld vload_half(size_t offset, const __local half *p);
-float __ovld vload_half(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Read sizeof (halfn) bytes of data from address
- * (p + (offset * n)). The data read is interpreted
- * as a halfn value. The halfn value read is
- * converted to a floatn value and the floatn
- * value is returned. The read address computed
- * as (p + (offset * n)) must be 16-bit aligned.
- */
-float2 __ovld vload_half2(size_t offset, const __constant half *p);
-float3 __ovld vload_half3(size_t offset, const __constant half *p);
-float4 __ovld vload_half4(size_t offset, const __constant half *p);
-float8 __ovld vload_half8(size_t offset, const __constant half *p);
-float16 __ovld vload_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float2 __ovld vload_half2(size_t offset, const half *p);
-float3 __ovld vload_half3(size_t offset, const half *p);
-float4 __ovld vload_half4(size_t offset, const half *p);
-float8 __ovld vload_half8(size_t offset, const half *p);
-float16 __ovld vload_half16(size_t offset, const half *p);
-#else
-float2 __ovld vload_half2(size_t offset, const __global half *p);
-float3 __ovld vload_half3(size_t offset, const __global half *p);
-float4 __ovld vload_half4(size_t offset, const __global half *p);
-float8 __ovld vload_half8(size_t offset, const __global half *p);
-float16 __ovld vload_half16(size_t offset, const __global half *p);
-float2 __ovld vload_half2(size_t offset, const __local half *p);
-float3 __ovld vload_half3(size_t offset, const __local half *p);
-float4 __ovld vload_half4(size_t offset, const __local half *p);
-float8 __ovld vload_half8(size_t offset, const __local half *p);
-float16 __ovld vload_half16(size_t offset, const __local half *p);
-float2 __ovld vload_half2(size_t offset, const __private half *p);
-float3 __ovld vload_half3(size_t offset, const __private half *p);
-float4 __ovld vload_half4(size_t offset, const __private half *p);
-float8 __ovld vload_half8(size_t offset, const __private half *p);
-float16 __ovld vload_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * The float value given by data is first
- * converted to a half value using the appropriate
- * rounding mode. The half value is then written
- * to address computed as (p + offset). The
- * address computed as (p + offset) must be 16-
- * bit aligned.
- * vstore_half use the current rounding mode.
- * The default current rounding mode is round to
- * nearest even.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld vstore_half(float data, size_t offset, half *p);
-void __ovld vstore_half_rte(float data, size_t offset, half *p);
-void __ovld vstore_half_rtz(float data, size_t offset, half *p);
-void __ovld vstore_half_rtp(float data, size_t offset, half *p);
-void __ovld vstore_half_rtn(float data, size_t offset, half *p);
-#ifdef cl_khr_fp64
-void __ovld vstore_half(double data, size_t offset, half *p);
-void __ovld vstore_half_rte(double data, size_t offset, half *p);
-void __ovld vstore_half_rtz(double data, size_t offset, half *p);
-void __ovld vstore_half_rtp(double data, size_t offset, half *p);
-void __ovld vstore_half_rtn(double data, size_t offset, half *p);
-#endif //cl_khr_fp64
-#else
-void __ovld vstore_half(float data, size_t offset, __global half *p);
-void __ovld vstore_half_rte(float data, size_t offset, __global half *p);
-void __ovld vstore_half_rtz(float data, size_t offset, __global half *p);
-void __ovld vstore_half_rtp(float data, size_t offset, __global half *p);
-void __ovld vstore_half_rtn(float data, size_t offset, __global half *p);
-void __ovld vstore_half(float data, size_t offset, __local half *p);
-void __ovld vstore_half_rte(float data, size_t offset, __local half *p);
-void __ovld vstore_half_rtz(float data, size_t offset, __local half *p);
-void __ovld vstore_half_rtp(float data, size_t offset, __local half *p);
-void __ovld vstore_half_rtn(float data, size_t offset, __local half *p);
-void __ovld vstore_half(float data, size_t offset, __private half *p);
-void __ovld vstore_half_rte(float data, size_t offset, __private half *p);
-void __ovld vstore_half_rtz(float data, size_t offset, __private half *p);
-void __ovld vstore_half_rtp(float data, size_t offset, __private half *p);
-void __ovld vstore_half_rtn(float data, size_t offset, __private half *p);
-#ifdef cl_khr_fp64
-void __ovld vstore_half(double data, size_t offset, __global half *p);
-void __ovld vstore_half_rte(double data, size_t offset, __global half *p);
-void __ovld vstore_half_rtz(double data, size_t offset, __global half *p);
-void __ovld vstore_half_rtp(double data, size_t offset, __global half *p);
-void __ovld vstore_half_rtn(double data, size_t offset, __global half *p);
-void __ovld vstore_half(double data, size_t offset, __local half *p);
-void __ovld vstore_half_rte(double data, size_t offset, __local half *p);
-void __ovld vstore_half_rtz(double data, size_t offset, __local half *p);
-void __ovld vstore_half_rtp(double data, size_t offset, __local half *p);
-void __ovld vstore_half_rtn(double data, size_t offset, __local half *p);
-void __ovld vstore_half(double data, size_t offset, __private half *p);
-void __ovld vstore_half_rte(double data, size_t offset, __private half *p);
-void __ovld vstore_half_rtz(double data, size_t offset, __private half *p);
-void __ovld vstore_half_rtp(double data, size_t offset, __private half *p);
-void __ovld vstore_half_rtn(double data, size_t offset, __private half *p);
-#endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * The floatn value given by data is converted to
- * a halfn value using the appropriate rounding
- * mode. The halfn value is then written to
- * address computed as (p + (offset * n)). The
- * address computed as (p + (offset * n)) must be
- * 16-bit aligned.
- * vstore_halfn uses the current rounding mode.
- * The default current rounding mode is round to
- * nearest even.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld vstore_half2(float2 data, size_t offset, half *p);
-void __ovld vstore_half3(float3 data, size_t offset, half *p);
-void __ovld vstore_half4(float4 data, size_t offset, half *p);
-void __ovld vstore_half8(float8 data, size_t offset, half *p);
-void __ovld vstore_half16(float16 data, size_t offset, half *p);
-void __ovld vstore_half2_rte(float2 data, size_t offset, half *p);
-void __ovld vstore_half3_rte(float3 data, size_t offset, half *p);
-void __ovld vstore_half4_rte(float4 data, size_t offset, half *p);
-void __ovld vstore_half8_rte(float8 data, size_t offset, half *p);
-void __ovld vstore_half16_rte(float16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtz(float2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtz(float3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtz(float4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtz(float8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtz(float16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtp(float2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtp(float3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtp(float4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtp(float8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtp(float16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtn(float2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtn(float3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtn(float4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtn(float8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtn(float16 data, size_t offset, half *p);
-#ifdef cl_khr_fp64
-void __ovld vstore_half2(double2 data, size_t offset, half *p);
-void __ovld vstore_half3(double3 data, size_t offset, half *p);
-void __ovld vstore_half4(double4 data, size_t offset, half *p);
-void __ovld vstore_half8(double8 data, size_t offset, half *p);
-void __ovld vstore_half16(double16 data, size_t offset, half *p);
-void __ovld vstore_half2_rte(double2 data, size_t offset, half *p);
-void __ovld vstore_half3_rte(double3 data, size_t offset, half *p);
-void __ovld vstore_half4_rte(double4 data, size_t offset, half *p);
-void __ovld vstore_half8_rte(double8 data, size_t offset, half *p);
-void __ovld vstore_half16_rte(double16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtz(double2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtz(double3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtz(double4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtz(double8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtz(double16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtp(double2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtp(double3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtp(double4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtp(double8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtp(double16 data, size_t offset, half *p);
-void __ovld vstore_half2_rtn(double2 data, size_t offset, half *p);
-void __ovld vstore_half3_rtn(double3 data, size_t offset, half *p);
-void __ovld vstore_half4_rtn(double4 data, size_t offset, half *p);
-void __ovld vstore_half8_rtn(double8 data, size_t offset, half *p);
-void __ovld vstore_half16_rtn(double16 data, size_t offset, half *p);
-#endif //cl_khr_fp64
-#else
-void __ovld vstore_half2(float2 data, size_t offset, __global half *p);
-void __ovld vstore_half3(float3 data, size_t offset, __global half *p);
-void __ovld vstore_half4(float4 data, size_t offset, __global half *p);
-void __ovld vstore_half8(float8 data, size_t offset, __global half *p);
-void __ovld vstore_half16(float16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rte(float2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rte(float3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rte(float4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rte(float8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rte(float16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtz(float2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtz(float3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtz(float4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtz(float8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtz(float16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtp(float2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtp(float3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtp(float4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtp(float8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtp(float16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtn(float2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtn(float3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtn(float4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtn(float8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtn(float16 data, size_t offset, __global half *p);
-void __ovld vstore_half2(float2 data, size_t offset, __local half *p);
-void __ovld vstore_half3(float3 data, size_t offset, __local half *p);
-void __ovld vstore_half4(float4 data, size_t offset, __local half *p);
-void __ovld vstore_half8(float8 data, size_t offset, __local half *p);
-void __ovld vstore_half16(float16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rte(float2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rte(float3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rte(float4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rte(float8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rte(float16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtz(float2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtz(float3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtz(float4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtz(float8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtz(float16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtp(float2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtp(float3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtp(float4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtp(float8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtp(float16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtn(float2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtn(float3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtn(float4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtn(float8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtn(float16 data, size_t offset, __local half *p);
-void __ovld vstore_half2(float2 data, size_t offset, __private half *p);
-void __ovld vstore_half3(float3 data, size_t offset, __private half *p);
-void __ovld vstore_half4(float4 data, size_t offset, __private half *p);
-void __ovld vstore_half8(float8 data, size_t offset, __private half *p);
-void __ovld vstore_half16(float16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rte(float2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rte(float3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rte(float4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rte(float8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rte(float16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtz(float2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtz(float3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtz(float4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtz(float8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtz(float16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtp(float2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtp(float3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtp(float4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtp(float8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtp(float16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtn(float2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtn(float3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtn(float4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtn(float8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtn(float16 data, size_t offset, __private half *p);
-#ifdef cl_khr_fp64
-void __ovld vstore_half2(double2 data, size_t offset, __global half *p);
-void __ovld vstore_half3(double3 data, size_t offset, __global half *p);
-void __ovld vstore_half4(double4 data, size_t offset, __global half *p);
-void __ovld vstore_half8(double8 data, size_t offset, __global half *p);
-void __ovld vstore_half16(double16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rte(double2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rte(double3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rte(double4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rte(double8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rte(double16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtz(double2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtz(double3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtz(double4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtz(double8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtz(double16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtp(double2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtp(double3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtp(double4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtp(double8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtp(double16 data, size_t offset, __global half *p);
-void __ovld vstore_half2_rtn(double2 data, size_t offset, __global half *p);
-void __ovld vstore_half3_rtn(double3 data, size_t offset, __global half *p);
-void __ovld vstore_half4_rtn(double4 data, size_t offset, __global half *p);
-void __ovld vstore_half8_rtn(double8 data, size_t offset, __global half *p);
-void __ovld vstore_half16_rtn(double16 data, size_t offset, __global half *p);
-void __ovld vstore_half2(double2 data, size_t offset, __local half *p);
-void __ovld vstore_half3(double3 data, size_t offset, __local half *p);
-void __ovld vstore_half4(double4 data, size_t offset, __local half *p);
-void __ovld vstore_half8(double8 data, size_t offset, __local half *p);
-void __ovld vstore_half16(double16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rte(double2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rte(double3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rte(double4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rte(double8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rte(double16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtz(double2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtz(double3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtz(double4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtz(double8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtz(double16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtp(double2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtp(double3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtp(double4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtp(double8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtp(double16 data, size_t offset, __local half *p);
-void __ovld vstore_half2_rtn(double2 data, size_t offset, __local half *p);
-void __ovld vstore_half3_rtn(double3 data, size_t offset, __local half *p);
-void __ovld vstore_half4_rtn(double4 data, size_t offset, __local half *p);
-void __ovld vstore_half8_rtn(double8 data, size_t offset, __local half *p);
-void __ovld vstore_half16_rtn(double16 data, size_t offset, __local half *p);
-void __ovld vstore_half2(double2 data, size_t offset, __private half *p);
-void __ovld vstore_half3(double3 data, size_t offset, __private half *p);
-void __ovld vstore_half4(double4 data, size_t offset, __private half *p);
-void __ovld vstore_half8(double8 data, size_t offset, __private half *p);
-void __ovld vstore_half16(double16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rte(double2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rte(double3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rte(double4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rte(double8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rte(double16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtz(double2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtz(double3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtz(double4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtz(double8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtz(double16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtp(double2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtp(double3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtp(double4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtp(double8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtp(double16 data, size_t offset, __private half *p);
-void __ovld vstore_half2_rtn(double2 data, size_t offset, __private half *p);
-void __ovld vstore_half3_rtn(double3 data, size_t offset, __private half *p);
-void __ovld vstore_half4_rtn(double4 data, size_t offset, __private half *p);
-void __ovld vstore_half8_rtn(double8 data, size_t offset, __private half *p);
-void __ovld vstore_half16_rtn(double16 data, size_t offset, __private half *p);
-#endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * For n = 1, 2, 4, 8 and 16 read sizeof (halfn)
- * bytes of data from address (p + (offset * n)).
- * The data read is interpreted as a halfn value.
- * The halfn value read is converted to a floatn
- * value and the floatn value is returned.
- * The address computed as (p + (offset * n))
- * must be aligned to sizeof (halfn) bytes.
- * For n = 3, vloada_half3 reads a half3 from
- * address (p + (offset * 4)) and returns a float3.
- * The address computed as (p + (offset * 4))
- * must be aligned to sizeof (half) * 4 bytes.
- */
-float __ovld vloada_half(size_t offset, const __constant half *p);
-float2 __ovld vloada_half2(size_t offset, const __constant half *p);
-float3 __ovld vloada_half3(size_t offset, const __constant half *p);
-float4 __ovld vloada_half4(size_t offset, const __constant half *p);
-float8 __ovld vloada_half8(size_t offset, const __constant half *p);
-float16 __ovld vloada_half16(size_t offset, const __constant half *p);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float __ovld vloada_half(size_t offset, const half *p);
-float2 __ovld vloada_half2(size_t offset, const half *p);
-float3 __ovld vloada_half3(size_t offset, const half *p);
-float4 __ovld vloada_half4(size_t offset, const half *p);
-float8 __ovld vloada_half8(size_t offset, const half *p);
-float16 __ovld vloada_half16(size_t offset, const half *p);
-#else
-float __ovld vloada_half(size_t offset, const __global half *p);
-float2 __ovld vloada_half2(size_t offset, const __global half *p);
-float3 __ovld vloada_half3(size_t offset, const __global half *p);
-float4 __ovld vloada_half4(size_t offset, const __global half *p);
-float8 __ovld vloada_half8(size_t offset, const __global half *p);
-float16 __ovld vloada_half16(size_t offset, const __global half *p);
-float __ovld vloada_half(size_t offset, const __local half *p);
-float2 __ovld vloada_half2(size_t offset, const __local half *p);
-float3 __ovld vloada_half3(size_t offset, const __local half *p);
-float4 __ovld vloada_half4(size_t offset, const __local half *p);
-float8 __ovld vloada_half8(size_t offset, const __local half *p);
-float16 __ovld vloada_half16(size_t offset, const __local half *p);
-float __ovld vloada_half(size_t offset, const __private half *p);
-float2 __ovld vloada_half2(size_t offset, const __private half *p);
-float3 __ovld vloada_half3(size_t offset, const __private half *p);
-float4 __ovld vloada_half4(size_t offset, const __private half *p);
-float8 __ovld vloada_half8(size_t offset, const __private half *p);
-float16 __ovld vloada_half16(size_t offset, const __private half *p);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * The floatn value given by data is converted to
- * a halfn value using the appropriate rounding
- * mode.
- * For n = 1, 2, 4, 8 and 16, the halfn value is
- * written to the address computed as (p + (offset
- * * n)). The address computed as (p + (offset *
- * n)) must be aligned to sizeof (halfn) bytes.
- * For n = 3, the half3 value is written to the
- * address computed as (p + (offset * 4)). The
- * address computed as (p + (offset * 4)) must be
- * aligned to sizeof (half) * 4 bytes.
- * vstorea_halfn uses the current rounding
- * mode. The default current rounding mode is
- * round to nearest even.
- */
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld vstorea_half(float data, size_t offset, half *p);
-void __ovld vstorea_half2(float2 data, size_t offset, half *p);
-void __ovld vstorea_half3(float3 data, size_t offset, half *p);
-void __ovld vstorea_half4(float4 data, size_t offset, half *p);
-void __ovld vstorea_half8(float8 data, size_t offset, half *p);
-void __ovld vstorea_half16(float16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rte(float data, size_t offset, half *p);
-void __ovld vstorea_half2_rte(float2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rte(float3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rte(float4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rte(float8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rte(float16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtz(float data, size_t offset, half *p);
-void __ovld vstorea_half2_rtz(float2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtz(float3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtz(float4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtz(float8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtz(float16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtp(float data, size_t offset, half *p);
-void __ovld vstorea_half2_rtp(float2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtp(float3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtp(float4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtp(float8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtp(float16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtn(float data, size_t offset, half *p);
-void __ovld vstorea_half2_rtn(float2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtn(float3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtn(float4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtn(float8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtn(float16 data, size_t offset, half *p);
-
-#ifdef cl_khr_fp64
-void __ovld vstorea_half(double data, size_t offset, half *p);
-void __ovld vstorea_half2(double2 data, size_t offset, half *p);
-void __ovld vstorea_half3(double3 data, size_t offset, half *p);
-void __ovld vstorea_half4(double4 data, size_t offset, half *p);
-void __ovld vstorea_half8(double8 data, size_t offset, half *p);
-void __ovld vstorea_half16(double16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rte(double data, size_t offset, half *p);
-void __ovld vstorea_half2_rte(double2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rte(double3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rte(double4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rte(double8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rte(double16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtz(double data, size_t offset, half *p);
-void __ovld vstorea_half2_rtz(double2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtz(double3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtz(double4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtz(double8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtz(double16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtp(double data, size_t offset, half *p);
-void __ovld vstorea_half2_rtp(double2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtp(double3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtp(double4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtp(double8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtp(double16 data, size_t offset, half *p);
-
-void __ovld vstorea_half_rtn(double data, size_t offset, half *p);
-void __ovld vstorea_half2_rtn(double2 data, size_t offset, half *p);
-void __ovld vstorea_half3_rtn(double3 data, size_t offset, half *p);
-void __ovld vstorea_half4_rtn(double4 data, size_t offset, half *p);
-void __ovld vstorea_half8_rtn(double8 data, size_t offset, half *p);
-void __ovld vstorea_half16_rtn(double16 data, size_t offset, half *p);
-#endif //cl_khr_fp64
-
-#else
-void __ovld vstorea_half(float data, size_t offset, __global half *p);
-void __ovld vstorea_half2(float2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3(float3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4(float4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8(float8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16(float16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rte(float data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rte(float2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rte(float3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rte(float4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rte(float8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rte(float16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtz(float data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtz(float2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtz(float3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtz(float4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtz(float8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtz(float16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtp(float data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtp(float2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtp(float3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtp(float4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtp(float8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtp(float16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtn(float data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtn(float2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtn(float3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtn(float4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtn(float8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtn(float16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half(float data, size_t offset, __local half *p);
-void __ovld vstorea_half2(float2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3(float3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4(float4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8(float8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16(float16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rte(float data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rte(float2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rte(float3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rte(float4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rte(float8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rte(float16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtz(float data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtz(float2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtz(float3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtz(float4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtz(float8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtz(float16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtp(float data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtp(float2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtp(float3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtp(float4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtp(float8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtp(float16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtn(float data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtn(float2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtn(float3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtn(float4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtn(float8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtn(float16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half(float data, size_t offset, __private half *p);
-void __ovld vstorea_half2(float2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3(float3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4(float4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8(float8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16(float16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rte(float data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rte(float2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rte(float3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rte(float4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rte(float8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rte(float16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtz(float data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtz(float2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rtz(float3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rtz(float4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rtz(float8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rtz(float16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtp(float data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtp(float2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rtp(float3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rtp(float4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rtp(float8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rtp(float16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtn(float data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtn(float2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rtn(float3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rtn(float4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rtn(float8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rtn(float16 data, size_t offset, __private half *p);
-
-#ifdef cl_khr_fp64
-void __ovld vstorea_half(double data, size_t offset, __global half *p);
-void __ovld vstorea_half2(double2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3(double3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4(double4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8(double8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16(double16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rte(double data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rte(double2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rte(double3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rte(double4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rte(double8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rte(double16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtz(double data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtz(double2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtz(double3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtz(double4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtz(double8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtz(double16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtp(double data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtp(double2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtp(double3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtp(double4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtp(double8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtp(double16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half_rtn(double data, size_t offset, __global half *p);
-void __ovld vstorea_half2_rtn(double2 data, size_t offset, __global half *p);
-void __ovld vstorea_half3_rtn(double3 data, size_t offset, __global half *p);
-void __ovld vstorea_half4_rtn(double4 data, size_t offset, __global half *p);
-void __ovld vstorea_half8_rtn(double8 data, size_t offset, __global half *p);
-void __ovld vstorea_half16_rtn(double16 data, size_t offset, __global half *p);
-
-void __ovld vstorea_half(double data, size_t offset, __local half *p);
-void __ovld vstorea_half2(double2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3(double3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4(double4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8(double8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16(double16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rte(double data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rte(double2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rte(double3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rte(double4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rte(double8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rte(double16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtz(double data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtz(double2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtz(double3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtz(double4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtz(double8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtz(double16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtp(double data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtp(double2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtp(double3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtp(double4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtp(double8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtp(double16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half_rtn(double data, size_t offset, __local half *p);
-void __ovld vstorea_half2_rtn(double2 data, size_t offset, __local half *p);
-void __ovld vstorea_half3_rtn(double3 data, size_t offset, __local half *p);
-void __ovld vstorea_half4_rtn(double4 data, size_t offset, __local half *p);
-void __ovld vstorea_half8_rtn(double8 data, size_t offset, __local half *p);
-void __ovld vstorea_half16_rtn(double16 data, size_t offset, __local half *p);
-
-void __ovld vstorea_half(double data, size_t offset, __private half *p);
-void __ovld vstorea_half2(double2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3(double3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4(double4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8(double8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16(double16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rte(double data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rte(double2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rte(double3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rte(double4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rte(double8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rte(double16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtz(double data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtz(double2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rtz(double3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rtz(double4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rtz(double8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rtz(double16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtp(double data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtp(double2 data, size_t offset, __private half *p);
-void __ovld vstorea_half3_rtp(double3 data, size_t offset, __private half *p);
-void __ovld vstorea_half4_rtp(double4 data, size_t offset, __private half *p);
-void __ovld vstorea_half8_rtp(double8 data, size_t offset, __private half *p);
-void __ovld vstorea_half16_rtp(double16 data, size_t offset, __private half *p);
-
-void __ovld vstorea_half_rtn(double data, size_t offset, __private half *p);
-void __ovld vstorea_half2_rtn(double2 data,size_t offset, __private half *p);
-void __ovld vstorea_half3_rtn(double3 data,size_t offset, __private half *p);
-void __ovld vstorea_half4_rtn(double4 data,size_t offset, __private half *p);
-void __ovld vstorea_half8_rtn(double8 data,size_t offset, __private half *p);
-void __ovld vstorea_half16_rtn(double16 data,size_t offset, __private half *p);
-#endif //cl_khr_fp64
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.8, v1.2 s6.12.8, v2.0 s6.13.8 - Synchronization Functions
-
-/**
- * All work-items in a work-group executing the kernel
- * on a processor must execute this function before any
- * are allowed to continue execution beyond the barrier.
- * This function must be encountered by all work-items in
- * a work-group executing the kernel.
- * If barrier is inside a conditional statement, then all
- * work-items must enter the conditional if any work-item
- * enters the conditional statement and executes the
- * barrier.
- * If barrer is inside a loop, all work-items must execute
- * the barrier for each iteration of the loop before any are
- * allowed to continue execution beyond the barrier.
- * The barrier function also queues a memory fence
- * (reads and writes) to ensure correct ordering of
- * memory operations to local or global memory.
- * The flags argument specifies the memory address space
- * and can be set to a combination of the following literal
- * values.
- * CLK_LOCAL_MEM_FENCE - The barrier function
- * will either flush any variables stored in local memory
- * or queue a memory fence to ensure correct ordering of
- * memory operations to local memory.
- * CLK_GLOBAL_MEM_FENCE - The barrier function
- * will queue a memory fence to ensure correct ordering
- * of memory operations to global memory. This can be
- * useful when work-items, for example, write to buffer or
- * image objects and then want to read the updated data.
- */
-
-void __ovld __conv barrier(cl_mem_fence_flags flags);
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld __conv work_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
-void __ovld __conv work_group_barrier(cl_mem_fence_flags flags);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.9, v1.2 s6.12.9 - Explicit Memory Fence Functions
-
-/**
- * Orders loads and stores of a work-item
- * executing a kernel. This means that loads
- * and stores preceding the mem_fence will
- * be committed to memory before any loads
- * and stores following the mem_fence.
- * The flags argument specifies the memory
- * address space and can be set to a
- * combination of the following literal
- * values:
- * CLK_LOCAL_MEM_FENCE
- * CLK_GLOBAL_MEM_FENCE.
- */
-void __ovld mem_fence(cl_mem_fence_flags flags);
-
-/**
- * Read memory barrier that orders only
- * loads.
- * The flags argument specifies the memory
- * address space and can be set to a
- * combination of the following literal
- * values:
- * CLK_LOCAL_MEM_FENCE
- * CLK_GLOBAL_MEM_FENCE.
- */
-void __ovld read_mem_fence(cl_mem_fence_flags flags);
-
-/**
- * Write memory barrier that orders only
- * stores.
- * The flags argument specifies the memory
- * address space and can be set to a
- * combination of the following literal
- * values:
- * CLK_LOCAL_MEM_FENCE
- * CLK_GLOBAL_MEM_FENCE.
- */
-void __ovld write_mem_fence(cl_mem_fence_flags flags);
-
-// OpenCL v2.0 s6.13.9 - Address Space Qualifier Functions
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-cl_mem_fence_flags __ovld get_fence(const void *ptr);
-cl_mem_fence_flags __ovld get_fence(void *ptr);
-
-/**
- * Builtin functions to_global, to_local, and to_private need to be declared as Clang builtin functions
- * and checked in Sema since they should be declared as
- *   addr gentype* to_addr (gentype*);
- * where gentype is builtin type or user defined type.
- */
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.10, v1.2 s6.12.10, v2.0 s6.13.10 - Async Copies from Global to Local Memory, Local to Global Memory, and Prefetch
-
-/**
- * event_t async_work_group_copy (
- * __global gentype *dst,
- * const __local gentype *src,
- * size_t num_elements,
- * event_t event)
- * Perform an async copy of num_elements
- * gentype elements from src to dst. The async
- * copy is performed by all work-items in a workgroup
- * and this built-in function must therefore
- * be encountered by all work-items in a workgroup
- * executing the kernel with the same
- * argument values; otherwise the results are
- * undefined.
- * Returns an event object that can be used by
- * wait_group_events to wait for the async copy
- * to finish. The event argument can also be used
- * to associate the async_work_group_copy with
- * a previous async copy allowing an event to be
- * shared by multiple async copies; otherwise event
- * should be zero.
- * If event argument is non-zero, the event object
- * supplied in event argument will be returned.
- * This function does not perform any implicit
- * synchronization of source data such as using a
- * barrier before performing the copy.
- */
-event_t __ovld async_work_group_copy(__local char *dst, const __global char *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar *dst, const __global uchar *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short *dst, const __global short *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort *dst, const __global ushort *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int *dst, const __global int *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint *dst, const __global uint *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long *dst, const __global long *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong *dst, const __global ulong *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float *dst, const __global float *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local char2 *dst, const __global char2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar2 *dst, const __global uchar2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short2 *dst, const __global short2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort2 *dst, const __global ushort2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int2 *dst, const __global int2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint2 *dst, const __global uint2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long2 *dst, const __global long2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong2 *dst, const __global ulong2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float2 *dst, const __global float2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local char3 *dst, const __global char3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar3 *dst, const __global uchar3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short3 *dst, const __global short3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort3 *dst, const __global ushort3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int3 *dst, const __global int3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint3 *dst, const __global uint3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long3 *dst, const __global long3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong3 *dst, const __global ulong3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float3 *dst, const __global float3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local char4 *dst, const __global char4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar4 *dst, const __global uchar4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short4 *dst, const __global short4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort4 *dst, const __global ushort4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int4 *dst, const __global int4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint4 *dst, const __global uint4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long4 *dst, const __global long4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong4 *dst, const __global ulong4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float4 *dst, const __global float4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local char8 *dst, const __global char8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar8 *dst, const __global uchar8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short8 *dst, const __global short8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort8 *dst, const __global ushort8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int8 *dst, const __global int8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint8 *dst, const __global uint8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long8 *dst, const __global long8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong8 *dst, const __global ulong8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float8 *dst, const __global float8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local char16 *dst, const __global char16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uchar16 *dst, const __global uchar16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local short16 *dst, const __global short16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ushort16 *dst, const __global ushort16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local int16 *dst, const __global int16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local uint16 *dst, const __global uint16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local long16 *dst, const __global long16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local ulong16 *dst, const __global ulong16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local float16 *dst, const __global float16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char *dst, const __local char *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar *dst, const __local uchar *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short *dst, const __local short *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort *dst, const __local ushort *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int *dst, const __local int *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint *dst, const __local uint *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long *dst, const __local long *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong *dst, const __local ulong *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float *dst, const __local float *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char2 *dst, const __local char2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar2 *dst, const __local uchar2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short2 *dst, const __local short2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort2 *dst, const __local ushort2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int2 *dst, const __local int2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint2 *dst, const __local uint2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long2 *dst, const __local long2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong2 *dst, const __local ulong2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float2 *dst, const __local float2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char3 *dst, const __local char3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar3 *dst, const __local uchar3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short3 *dst, const __local short3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort3 *dst, const __local ushort3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int3 *dst, const __local int3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint3 *dst, const __local uint3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long3 *dst, const __local long3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong3 *dst, const __local ulong3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float3 *dst, const __local float3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char4 *dst, const __local char4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar4 *dst, const __local uchar4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short4 *dst, const __local short4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort4 *dst, const __local ushort4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int4 *dst, const __local int4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint4 *dst, const __local uint4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long4 *dst, const __local long4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong4 *dst, const __local ulong4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float4 *dst, const __local float4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char8 *dst, const __local char8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar8 *dst, const __local uchar8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short8 *dst, const __local short8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort8 *dst, const __local ushort8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int8 *dst, const __local int8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint8 *dst, const __local uint8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long8 *dst, const __local long8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong8 *dst, const __local ulong8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float8 *dst, const __local float8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global char16 *dst, const __local char16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uchar16 *dst, const __local uchar16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global short16 *dst, const __local short16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ushort16 *dst, const __local ushort16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global int16 *dst, const __local int16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global uint16 *dst, const __local uint16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global long16 *dst, const __local long16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global ulong16 *dst, const __local ulong16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global float16 *dst, const __local float16 *src, size_t num_elements, event_t event);
-#ifdef cl_khr_fp64
-event_t __ovld async_work_group_copy(__local double *dst, const __global double *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local double2 *dst, const __global double2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local double3 *dst, const __global double3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local double4 *dst, const __global double4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local double8 *dst, const __global double8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local double16 *dst, const __global double16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double *dst, const __local double *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double2 *dst, const __local double2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double3 *dst, const __local double3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double4 *dst, const __local double4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double8 *dst, const __local double8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global double16 *dst, const __local double16 *src, size_t num_elements, event_t event);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-event_t __ovld async_work_group_copy(__local half *dst, const __global half *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local half2 *dst, const __global half2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local half3 *dst, const __global half3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local half4 *dst, const __global half4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local half8 *dst, const __global half8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__local half16 *dst, const __global half16 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half *dst, const __local half *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half2 *dst, const __local half2 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half3 *dst, const __local half3 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half4 *dst, const __local half4 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half8 *dst, const __local half8 *src, size_t num_elements, event_t event);
-event_t __ovld async_work_group_copy(__global half16 *dst, const __local half16 *src, size_t num_elements, event_t event);
-#endif //cl_khr_fp16
-
-/**
- * Perform an async gather of num_elements
- * gentype elements from src to dst. The
- * src_stride is the stride in elements for each
- * gentype element read from src. The dst_stride
- * is the stride in elements for each gentype
- * element written to dst. The async gather is
- * performed by all work-items in a work-group.
- * This built-in function must therefore be
- * encountered by all work-items in a work-group
- * executing the kernel with the same argument
- * values; otherwise the results are undefined.
- * Returns an event object that can be used by
- * wait_group_events to wait for the async copy
- * to finish. The event argument can also be used
- * to associate the
- * async_work_group_strided_copy with a
- * previous async copy allowing an event to be
- * shared by multiple async copies; otherwise event
- * should be zero.
- * If event argument is non-zero, the event object
- * supplied in event argument will be returned.
- * This function does not perform any implicit
- * synchronization of source data such as using a
- * barrier before performing the copy.
- */
-event_t __ovld async_work_group_strided_copy(__local char *dst, const __global char *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar *dst, const __global uchar *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short *dst, const __global short *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort *dst, const __global ushort *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int *dst, const __global int *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint *dst, const __global uint *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long *dst, const __global long *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong *dst, const __global ulong *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float *dst, const __global float *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local char2 *dst, const __global char2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar2 *dst, const __global uchar2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short2 *dst, const __global short2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort2 *dst, const __global ushort2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int2 *dst, const __global int2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint2 *dst, const __global uint2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long2 *dst, const __global long2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong2 *dst, const __global ulong2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float2 *dst, const __global float2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local char3 *dst, const __global char3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar3 *dst, const __global uchar3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short3 *dst, const __global short3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort3 *dst, const __global ushort3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int3 *dst, const __global int3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint3 *dst, const __global uint3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long3 *dst, const __global long3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong3 *dst, const __global ulong3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float3 *dst, const __global float3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local char4 *dst, const __global char4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar4 *dst, const __global uchar4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short4 *dst, const __global short4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort4 *dst, const __global ushort4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int4 *dst, const __global int4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint4 *dst, const __global uint4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long4 *dst, const __global long4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong4 *dst, const __global ulong4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float4 *dst, const __global float4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local char8 *dst, const __global char8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar8 *dst, const __global uchar8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short8 *dst, const __global short8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort8 *dst, const __global ushort8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int8 *dst, const __global int8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint8 *dst, const __global uint8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long8 *dst, const __global long8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong8 *dst, const __global ulong8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float8 *dst, const __global float8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local char16 *dst, const __global char16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uchar16 *dst, const __global uchar16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local short16 *dst, const __global short16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ushort16 *dst, const __global ushort16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local int16 *dst, const __global int16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local uint16 *dst, const __global uint16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local long16 *dst, const __global long16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local ulong16 *dst, const __global ulong16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local float16 *dst, const __global float16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char *dst, const __local char *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar *dst, const __local uchar *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short *dst, const __local short *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort *dst, const __local ushort *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int *dst, const __local int *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint *dst, const __local uint *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long *dst, const __local long *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong *dst, const __local ulong *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float *dst, const __local float *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char2 *dst, const __local char2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar2 *dst, const __local uchar2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short2 *dst, const __local short2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort2 *dst, const __local ushort2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int2 *dst, const __local int2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint2 *dst, const __local uint2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long2 *dst, const __local long2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong2 *dst, const __local ulong2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float2 *dst, const __local float2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char3 *dst, const __local char3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar3 *dst, const __local uchar3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short3 *dst, const __local short3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort3 *dst, const __local ushort3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int3 *dst, const __local int3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint3 *dst, const __local uint3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long3 *dst, const __local long3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong3 *dst, const __local ulong3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float3 *dst, const __local float3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char4 *dst, const __local char4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar4 *dst, const __local uchar4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short4 *dst, const __local short4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort4 *dst, const __local ushort4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int4 *dst, const __local int4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint4 *dst, const __local uint4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long4 *dst, const __local long4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong4 *dst, const __local ulong4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float4 *dst, const __local float4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char8 *dst, const __local char8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar8 *dst, const __local uchar8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short8 *dst, const __local short8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort8 *dst, const __local ushort8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int8 *dst, const __local int8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint8 *dst, const __local uint8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long8 *dst, const __local long8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong8 *dst, const __local ulong8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float8 *dst, const __local float8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global char16 *dst, const __local char16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uchar16 *dst, const __local uchar16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global short16 *dst, const __local short16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ushort16 *dst, const __local ushort16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global int16 *dst, const __local int16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global uint16 *dst, const __local uint16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global long16 *dst, const __local long16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global ulong16 *dst, const __local ulong16 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global float16 *dst, const __local float16 *src, size_t num_elements, size_t dst_stride, event_t event);
-#ifdef cl_khr_fp64
-event_t __ovld async_work_group_strided_copy(__local double *dst, const __global double *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local double2 *dst, const __global double2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local double3 *dst, const __global double3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local double4 *dst, const __global double4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local double8 *dst, const __global double8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local double16 *dst, const __global double16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double *dst, const __local double *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double2 *dst, const __local double2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double3 *dst, const __local double3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double4 *dst, const __local double4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double8 *dst, const __local double8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global double16 *dst, const __local double16 *src, size_t num_elements, size_t dst_stride, event_t event);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-event_t __ovld async_work_group_strided_copy(__local half *dst, const __global half *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local half2 *dst, const __global half2 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local half3 *dst, const __global half3 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local half4 *dst, const __global half4 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local half8 *dst, const __global half8 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__local half16 *dst, const __global half16 *src, size_t num_elements, size_t src_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half *dst, const __local half *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half2 *dst, const __local half2 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half3 *dst, const __local half3 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half4 *dst, const __local half4 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half8 *dst, const __local half8 *src, size_t num_elements, size_t dst_stride, event_t event);
-event_t __ovld async_work_group_strided_copy(__global half16 *dst, const __local half16 *src, size_t num_elements, size_t dst_stride, event_t event);
-#endif //cl_khr_fp16
-
-/**
- * Wait for events that identify the
- * async_work_group_copy operations to
- * complete. The event objects specified in
- * event_list will be released after the wait is
- * performed.
- * This function must be encountered by all workitems
- * in a work-group executing the kernel with
- * the same num_events and event objects specified
- * in event_list; otherwise the results are undefined.
- */
-void __ovld wait_group_events(int num_events, event_t *event_list);
-
-/**
- * Prefetch num_elements * sizeof(gentype)
- * bytes into the global cache. The prefetch
- * instruction is applied to a work-item in a workgroup
- * and does not affect the functional
- * behavior of the kernel.
- */
-void __ovld prefetch(const __global char *p, size_t num_elements);
-void __ovld prefetch(const __global uchar *p, size_t num_elements);
-void __ovld prefetch(const __global short *p, size_t num_elements);
-void __ovld prefetch(const __global ushort *p, size_t num_elements);
-void __ovld prefetch(const __global int *p, size_t num_elements);
-void __ovld prefetch(const __global uint *p, size_t num_elements);
-void __ovld prefetch(const __global long *p, size_t num_elements);
-void __ovld prefetch(const __global ulong *p, size_t num_elements);
-void __ovld prefetch(const __global float *p, size_t num_elements);
-void __ovld prefetch(const __global char2 *p, size_t num_elements);
-void __ovld prefetch(const __global uchar2 *p, size_t num_elements);
-void __ovld prefetch(const __global short2 *p, size_t num_elements);
-void __ovld prefetch(const __global ushort2 *p, size_t num_elements);
-void __ovld prefetch(const __global int2 *p, size_t num_elements);
-void __ovld prefetch(const __global uint2 *p, size_t num_elements);
-void __ovld prefetch(const __global long2 *p, size_t num_elements);
-void __ovld prefetch(const __global ulong2 *p, size_t num_elements);
-void __ovld prefetch(const __global float2 *p, size_t num_elements);
-void __ovld prefetch(const __global char3 *p, size_t num_elements);
-void __ovld prefetch(const __global uchar3 *p, size_t num_elements);
-void __ovld prefetch(const __global short3 *p, size_t num_elements);
-void __ovld prefetch(const __global ushort3 *p, size_t num_elements);
-void __ovld prefetch(const __global int3 *p, size_t num_elements);
-void __ovld prefetch(const __global uint3 *p, size_t num_elements);
-void __ovld prefetch(const __global long3 *p, size_t num_elements);
-void __ovld prefetch(const __global ulong3 *p, size_t num_elements);
-void __ovld prefetch(const __global float3 *p, size_t num_elements);
-void __ovld prefetch(const __global char4 *p, size_t num_elements);
-void __ovld prefetch(const __global uchar4 *p, size_t num_elements);
-void __ovld prefetch(const __global short4 *p, size_t num_elements);
-void __ovld prefetch(const __global ushort4 *p, size_t num_elements);
-void __ovld prefetch(const __global int4 *p, size_t num_elements);
-void __ovld prefetch(const __global uint4 *p, size_t num_elements);
-void __ovld prefetch(const __global long4 *p, size_t num_elements);
-void __ovld prefetch(const __global ulong4 *p, size_t num_elements);
-void __ovld prefetch(const __global float4 *p, size_t num_elements);
-void __ovld prefetch(const __global char8 *p, size_t num_elements);
-void __ovld prefetch(const __global uchar8 *p, size_t num_elements);
-void __ovld prefetch(const __global short8 *p, size_t num_elements);
-void __ovld prefetch(const __global ushort8 *p, size_t num_elements);
-void __ovld prefetch(const __global int8 *p, size_t num_elements);
-void __ovld prefetch(const __global uint8 *p, size_t num_elements);
-void __ovld prefetch(const __global long8 *p, size_t num_elements);
-void __ovld prefetch(const __global ulong8 *p, size_t num_elements);
-void __ovld prefetch(const __global float8 *p, size_t num_elements);
-void __ovld prefetch(const __global char16 *p, size_t num_elements);
-void __ovld prefetch(const __global uchar16 *p, size_t num_elements);
-void __ovld prefetch(const __global short16 *p, size_t num_elements);
-void __ovld prefetch(const __global ushort16 *p, size_t num_elements);
-void __ovld prefetch(const __global int16 *p, size_t num_elements);
-void __ovld prefetch(const __global uint16 *p, size_t num_elements);
-void __ovld prefetch(const __global long16 *p, size_t num_elements);
-void __ovld prefetch(const __global ulong16 *p, size_t num_elements);
-void __ovld prefetch(const __global float16 *p, size_t num_elements);
-#ifdef cl_khr_fp64
-void __ovld prefetch(const __global double *p, size_t num_elements);
-void __ovld prefetch(const __global double2 *p, size_t num_elements);
-void __ovld prefetch(const __global double3 *p, size_t num_elements);
-void __ovld prefetch(const __global double4 *p, size_t num_elements);
-void __ovld prefetch(const __global double8 *p, size_t num_elements);
-void __ovld prefetch(const __global double16 *p, size_t num_elements);
-#endif //cl_khr_fp64
-#ifdef cl_khr_fp16
-void __ovld prefetch(const __global half *p, size_t num_elements);
-void __ovld prefetch(const __global half2 *p, size_t num_elements);
-void __ovld prefetch(const __global half3 *p, size_t num_elements);
-void __ovld prefetch(const __global half4 *p, size_t num_elements);
-void __ovld prefetch(const __global half8 *p, size_t num_elements);
-void __ovld prefetch(const __global half16 *p, size_t num_elements);
-#endif // cl_khr_fp16
-
-// OpenCL v1.1 s6.11.1, v1.2 s6.12.11 - Atomic Functions
-
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#pragma OPENCL EXTENSION cl_khr_int64_base_atomics : enable
-#pragma OPENCL EXTENSION cl_khr_int64_extended_atomics : enable
-#endif
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old + val) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_add(volatile __global int *p, int val);
-unsigned int __ovld atomic_add(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_add(volatile __local int *p, int val);
-unsigned int __ovld atomic_add(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_add(volatile int *p, int val);
-unsigned int __ovld atomic_add(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_add(volatile __global int *p, int val);
-unsigned int __ovld atom_add(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_add(volatile __local int *p, int val);
-unsigned int __ovld atom_add(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_add(volatile __global long *p, long val);
-unsigned long __ovld atom_add(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_add(volatile __local long *p, long val);
-unsigned long __ovld atom_add(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old) stored at location pointed by p.
- * Compute (old - val) and store result at location pointed by p. The function
- * returns old.
- */
-int __ovld atomic_sub(volatile __global int *p, int val);
-unsigned int __ovld atomic_sub(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_sub(volatile __local int *p, int val);
-unsigned int __ovld atomic_sub(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_sub(volatile int *p, int val);
-unsigned int __ovld atomic_sub(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_sub(volatile __global int *p, int val);
-unsigned int __ovld atom_sub(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_sub(volatile __local int *p, int val);
-unsigned int __ovld atom_sub(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_sub(volatile __global long *p, long val);
-unsigned long __ovld atom_sub(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_sub(volatile __local long *p, long val);
-unsigned long __ovld atom_sub(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Swaps the old value stored at location p
- * with new value given by val. Returns old
- * value.
- */
-int __ovld atomic_xchg(volatile __global int *p, int val);
-unsigned int __ovld atomic_xchg(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_xchg(volatile __local int *p, int val);
-unsigned int __ovld atomic_xchg(volatile __local unsigned int *p, unsigned int val);
-float __ovld atomic_xchg(volatile __global float *p, float val);
-float __ovld atomic_xchg(volatile __local float *p, float val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_xchg(volatile int *p, int val);
-unsigned int __ovld atomic_xchg(volatile unsigned int *p, unsigned int val);
-float __ovld atomic_xchg(volatile float *p, float val);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_xchg(volatile __global int *p, int val);
-unsigned int __ovld atom_xchg(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_xchg(volatile __local int *p, int val);
-unsigned int __ovld atom_xchg(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_xchg(volatile __global long *p, long val);
-long __ovld atom_xchg(volatile __local long *p, long val);
-unsigned long __ovld atom_xchg(volatile __global unsigned long *p, unsigned long val);
-unsigned long __ovld atom_xchg(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old + 1) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_inc(volatile __global int *p);
-unsigned int __ovld atomic_inc(volatile __global unsigned int *p);
-int __ovld atomic_inc(volatile __local int *p);
-unsigned int __ovld atomic_inc(volatile __local unsigned int *p);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_inc(volatile int *p);
-unsigned int __ovld atomic_inc(volatile unsigned int *p);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_inc(volatile __global int *p);
-unsigned int __ovld atom_inc(volatile __global unsigned int *p);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_inc(volatile __local int *p);
-unsigned int __ovld atom_inc(volatile __local unsigned int *p);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_inc(volatile __global long *p);
-unsigned long __ovld atom_inc(volatile __global unsigned long *p);
-long __ovld atom_inc(volatile __local long *p);
-unsigned long __ovld atom_inc(volatile __local unsigned long *p);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old - 1) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_dec(volatile __global int *p);
-unsigned int __ovld atomic_dec(volatile __global unsigned int *p);
-int __ovld atomic_dec(volatile __local int *p);
-unsigned int __ovld atomic_dec(volatile __local unsigned int *p);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_dec(volatile int *p);
-unsigned int __ovld atomic_dec(volatile unsigned int *p);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_dec(volatile __global int *p);
-unsigned int __ovld atom_dec(volatile __global unsigned int *p);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_dec(volatile __local int *p);
-unsigned int __ovld atom_dec(volatile __local unsigned int *p);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_dec(volatile __global long *p);
-unsigned long __ovld atom_dec(volatile __global unsigned long *p);
-long __ovld atom_dec(volatile __local long *p);
-unsigned long __ovld atom_dec(volatile __local unsigned long *p);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old == cmp) ? val : old and store result at
- * location pointed by p. The function
- * returns old.
- */
-int __ovld atomic_cmpxchg(volatile __global int *p, int cmp, int val);
-unsigned int __ovld atomic_cmpxchg(volatile __global unsigned int *p, unsigned int cmp, unsigned int val);
-int __ovld atomic_cmpxchg(volatile __local int *p, int cmp, int val);
-unsigned int __ovld atomic_cmpxchg(volatile __local unsigned int *p, unsigned int cmp, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_cmpxchg(volatile int *p, int cmp, int val);
-unsigned int __ovld atomic_cmpxchg(volatile unsigned int *p, unsigned int cmp, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_base_atomics)
-int __ovld atom_cmpxchg(volatile __global int *p, int cmp, int val);
-unsigned int __ovld atom_cmpxchg(volatile __global unsigned int *p, unsigned int cmp, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_base_atomics)
-int __ovld atom_cmpxchg(volatile __local int *p, int cmp, int val);
-unsigned int __ovld atom_cmpxchg(volatile __local unsigned int *p, unsigned int cmp, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_base_atomics)
-long __ovld atom_cmpxchg(volatile __global long *p, long cmp, long val);
-unsigned long __ovld atom_cmpxchg(volatile __global unsigned long *p, unsigned long cmp, unsigned long val);
-long __ovld atom_cmpxchg(volatile __local long *p, long cmp, long val);
-unsigned long __ovld atom_cmpxchg(volatile __local unsigned long *p, unsigned long cmp, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * min(old, val) and store minimum value at
- * location pointed by p. The function
- * returns old.
- */
-int __ovld atomic_min(volatile __global int *p, int val);
-unsigned int __ovld atomic_min(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_min(volatile __local int *p, int val);
-unsigned int __ovld atomic_min(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_min(volatile int *p, int val);
-unsigned int __ovld atomic_min(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_extended_atomics)
-int __ovld atom_min(volatile __global int *p, int val);
-unsigned int __ovld atom_min(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_extended_atomics)
-int __ovld atom_min(volatile __local int *p, int val);
-unsigned int __ovld atom_min(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_extended_atomics)
-long __ovld atom_min(volatile __global long *p, long val);
-unsigned long __ovld atom_min(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_min(volatile __local long *p, long val);
-unsigned long __ovld atom_min(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * max(old, val) and store maximum value at
- * location pointed by p. The function
- * returns old.
- */
-int __ovld atomic_max(volatile __global int *p, int val);
-unsigned int __ovld atomic_max(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_max(volatile __local int *p, int val);
-unsigned int __ovld atomic_max(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_max(volatile int *p, int val);
-unsigned int __ovld atomic_max(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_extended_atomics)
-int __ovld atom_max(volatile __global int *p, int val);
-unsigned int __ovld atom_max(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_extended_atomics)
-int __ovld atom_max(volatile __local int *p, int val);
-unsigned int __ovld atom_max(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_extended_atomics)
-long __ovld atom_max(volatile __global long *p, long val);
-unsigned long __ovld atom_max(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_max(volatile __local long *p, long val);
-unsigned long __ovld atom_max(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old & val) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_and(volatile __global int *p, int val);
-unsigned int __ovld atomic_and(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_and(volatile __local int *p, int val);
-unsigned int __ovld atomic_and(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_and(volatile int *p, int val);
-unsigned int __ovld atomic_and(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_extended_atomics)
-int __ovld atom_and(volatile __global int *p, int val);
-unsigned int __ovld atom_and(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_extended_atomics)
-int __ovld atom_and(volatile __local int *p, int val);
-unsigned int __ovld atom_and(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_extended_atomics)
-long __ovld atom_and(volatile __global long *p, long val);
-unsigned long __ovld atom_and(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_and(volatile __local long *p, long val);
-unsigned long __ovld atom_and(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old | val) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_or(volatile __global int *p, int val);
-unsigned int __ovld atomic_or(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_or(volatile __local int *p, int val);
-unsigned int __ovld atomic_or(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_or(volatile int *p, int val);
-unsigned int __ovld atomic_or(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_extended_atomics)
-int __ovld atom_or(volatile __global int *p, int val);
-unsigned int __ovld atom_or(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_extended_atomics)
-int __ovld atom_or(volatile __local int *p, int val);
-unsigned int __ovld atom_or(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_extended_atomics)
-long __ovld atom_or(volatile __global long *p, long val);
-unsigned long __ovld atom_or(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_or(volatile __local long *p, long val);
-unsigned long __ovld atom_or(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-/**
- * Read the 32-bit value (referred to as old)
- * stored at location pointed by p. Compute
- * (old ^ val) and store result at location
- * pointed by p. The function returns old.
- */
-int __ovld atomic_xor(volatile __global int *p, int val);
-unsigned int __ovld atomic_xor(volatile __global unsigned int *p, unsigned int val);
-int __ovld atomic_xor(volatile __local int *p, int val);
-unsigned int __ovld atomic_xor(volatile __local unsigned int *p, unsigned int val);
-#ifdef __OPENCL_CPP_VERSION__
-int __ovld atomic_xor(volatile int *p, int val);
-unsigned int __ovld atomic_xor(volatile unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_global_int32_extended_atomics)
-int __ovld atom_xor(volatile __global int *p, int val);
-unsigned int __ovld atom_xor(volatile __global unsigned int *p, unsigned int val);
-#endif
-#if defined(cl_khr_local_int32_extended_atomics)
-int __ovld atom_xor(volatile __local int *p, int val);
-unsigned int __ovld atom_xor(volatile __local unsigned int *p, unsigned int val);
-#endif
-
-#if defined(cl_khr_int64_extended_atomics)
-long __ovld atom_xor(volatile __global long *p, long val);
-unsigned long __ovld atom_xor(volatile __global unsigned long *p, unsigned long val);
-long __ovld atom_xor(volatile __local long *p, long val);
-unsigned long __ovld atom_xor(volatile __local unsigned long *p, unsigned long val);
-#endif
-
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#pragma OPENCL EXTENSION cl_khr_int64_base_atomics : disable
-#pragma OPENCL EXTENSION cl_khr_int64_extended_atomics : disable
-#endif
-
-// OpenCL v2.0 s6.13.11 - Atomics Functions
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// double atomics support requires extensions cl_khr_int64_base_atomics and cl_khr_int64_extended_atomics
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#pragma OPENCL EXTENSION cl_khr_int64_base_atomics : enable
-#pragma OPENCL EXTENSION cl_khr_int64_extended_atomics : enable
-#endif
-
-// atomic_init()
-void __ovld atomic_init(volatile atomic_int *object, int value);
-void __ovld atomic_init(volatile atomic_uint *object, uint value);
-void __ovld atomic_init(volatile atomic_float *object, float value);
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-void __ovld atomic_init(volatile atomic_long *object, long value);
-void __ovld atomic_init(volatile atomic_ulong *object, ulong value);
-#ifdef cl_khr_fp64
-void __ovld atomic_init(volatile atomic_double *object, double value);
-#endif //cl_khr_fp64
-#endif
-
-// atomic_work_item_fence()
-void __ovld atomic_work_item_fence(cl_mem_fence_flags flags, memory_order order, memory_scope scope);
-
-// atomic_fetch()
-
-int __ovld atomic_fetch_add(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_add_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_add_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_add(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_add_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_add_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_sub(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_sub_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_sub_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_sub(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_sub_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_sub_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_or(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_or_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_or_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_or(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_or_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_or_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_xor(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_xor_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_xor_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_xor(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_xor_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_xor_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_and(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_and_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_and_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_and(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_and_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_and_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_min(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_min_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_min_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_min(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_min_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_min_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_min(volatile atomic_uint *object, int operand);
-uint __ovld atomic_fetch_min_explicit(volatile atomic_uint *object, int operand, memory_order order);
-uint __ovld atomic_fetch_min_explicit(volatile atomic_uint *object, int operand, memory_order order, memory_scope scope);
-int __ovld atomic_fetch_max(volatile atomic_int *object, int operand);
-int __ovld atomic_fetch_max_explicit(volatile atomic_int *object, int operand, memory_order order);
-int __ovld atomic_fetch_max_explicit(volatile atomic_int *object, int operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_max(volatile atomic_uint *object, uint operand);
-uint __ovld atomic_fetch_max_explicit(volatile atomic_uint *object, uint operand, memory_order order);
-uint __ovld atomic_fetch_max_explicit(volatile atomic_uint *object, uint operand, memory_order order, memory_scope scope);
-uint __ovld atomic_fetch_max(volatile atomic_uint *object, int operand);
-uint __ovld atomic_fetch_max_explicit(volatile atomic_uint *object, int operand, memory_order order);
-uint __ovld atomic_fetch_max_explicit(volatile atomic_uint *object, int operand, memory_order order, memory_scope scope);
-
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-long __ovld atomic_fetch_add(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_add_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_add_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_add(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_add_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_add_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_sub(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_sub_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_sub_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_sub(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_sub_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_sub_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_or(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_or_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_or_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_or(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_or_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_or_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_xor(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_xor_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_xor_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_xor(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_xor_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_xor_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_and(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_and_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_and_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_and(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_and_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_and_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_min(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_min_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_min_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_min(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_min_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_min_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_min(volatile atomic_ulong *object, long operand);
-ulong __ovld atomic_fetch_min_explicit(volatile atomic_ulong *object, long operand, memory_order order);
-ulong __ovld atomic_fetch_min_explicit(volatile atomic_ulong *object, long operand, memory_order order, memory_scope scope);
-long __ovld atomic_fetch_max(volatile atomic_long *object, long operand);
-long __ovld atomic_fetch_max_explicit(volatile atomic_long *object, long operand, memory_order order);
-long __ovld atomic_fetch_max_explicit(volatile atomic_long *object, long operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_max(volatile atomic_ulong *object, ulong operand);
-ulong __ovld atomic_fetch_max_explicit(volatile atomic_ulong *object, ulong operand, memory_order order);
-ulong __ovld atomic_fetch_max_explicit(volatile atomic_ulong *object, ulong operand, memory_order order, memory_scope scope);
-ulong __ovld atomic_fetch_max(volatile atomic_ulong *object, long operand);
-ulong __ovld atomic_fetch_max_explicit(volatile atomic_ulong *object, long operand, memory_order order);
-ulong __ovld atomic_fetch_max_explicit(volatile atomic_ulong *object, long operand, memory_order order, memory_scope scope);
-#endif //defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-
-// OpenCL v2.0 s6.13.11.7.5:
-// add/sub: atomic type argument can be uintptr_t/intptr_t, value type argument can be ptrdiff_t.
-// or/xor/and/min/max: atomic type argument can be intptr_t/uintptr_t, value type argument can be intptr_t/uintptr_t.
-
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-uintptr_t __ovld atomic_fetch_add(volatile atomic_uintptr_t *object, ptrdiff_t operand);
-uintptr_t __ovld atomic_fetch_add_explicit(volatile atomic_uintptr_t *object, ptrdiff_t operand, memory_order order);
-uintptr_t __ovld atomic_fetch_add_explicit(volatile atomic_uintptr_t *object, ptrdiff_t operand, memory_order order, memory_scope scope);
-uintptr_t __ovld atomic_fetch_sub(volatile atomic_uintptr_t *object, ptrdiff_t operand);
-uintptr_t __ovld atomic_fetch_sub_explicit(volatile atomic_uintptr_t *object, ptrdiff_t operand, memory_order order);
-uintptr_t __ovld atomic_fetch_sub_explicit(volatile atomic_uintptr_t *object, ptrdiff_t operand, memory_order order, memory_scope scope);
-
-uintptr_t __ovld atomic_fetch_or(volatile atomic_uintptr_t *object, intptr_t operand);
-uintptr_t __ovld atomic_fetch_or_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order);
-uintptr_t __ovld atomic_fetch_or_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order, memory_scope scope);
-uintptr_t __ovld atomic_fetch_xor(volatile atomic_uintptr_t *object, intptr_t operand);
-uintptr_t __ovld atomic_fetch_xor_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order);
-uintptr_t __ovld atomic_fetch_xor_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order, memory_scope scope);
-uintptr_t __ovld atomic_fetch_and(volatile atomic_uintptr_t *object, intptr_t operand);
-uintptr_t __ovld atomic_fetch_and_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order);
-uintptr_t __ovld atomic_fetch_and_explicit(volatile atomic_uintptr_t *object, intptr_t operand, memory_order order, memory_scope scope);
-uintptr_t __ovld atomic_fetch_min(volatile atomic_uintptr_t *object, intptr_t opermax);
-uintptr_t __ovld atomic_fetch_min_explicit(volatile atomic_uintptr_t *object, intptr_t opermax, memory_order minder);
-uintptr_t __ovld atomic_fetch_min_explicit(volatile atomic_uintptr_t *object, intptr_t opermax, memory_order minder, memory_scope scope);
-uintptr_t __ovld atomic_fetch_max(volatile atomic_uintptr_t *object, intptr_t opermax);
-uintptr_t __ovld atomic_fetch_max_explicit(volatile atomic_uintptr_t *object, intptr_t opermax, memory_order minder);
-uintptr_t __ovld atomic_fetch_max_explicit(volatile atomic_uintptr_t *object, intptr_t opermax, memory_order minder, memory_scope scope);
-
-intptr_t __ovld atomic_fetch_or(volatile atomic_intptr_t *object, uintptr_t operand);
-intptr_t __ovld atomic_fetch_or_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order);
-intptr_t __ovld atomic_fetch_or_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order, memory_scope scope);
-intptr_t __ovld atomic_fetch_xor(volatile atomic_intptr_t *object, uintptr_t operand);
-intptr_t __ovld atomic_fetch_xor_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order);
-intptr_t __ovld atomic_fetch_xor_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order, memory_scope scope);
-intptr_t __ovld atomic_fetch_and(volatile atomic_intptr_t *object, uintptr_t operand);
-intptr_t __ovld atomic_fetch_and_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order);
-intptr_t __ovld atomic_fetch_and_explicit(volatile atomic_intptr_t *object, uintptr_t operand, memory_order order, memory_scope scope);
-intptr_t __ovld atomic_fetch_min(volatile atomic_intptr_t *object, uintptr_t opermax);
-intptr_t __ovld atomic_fetch_min_explicit(volatile atomic_intptr_t *object, uintptr_t opermax, memory_order minder);
-intptr_t __ovld atomic_fetch_min_explicit(volatile atomic_intptr_t *object, uintptr_t opermax, memory_order minder, memory_scope scope);
-intptr_t __ovld atomic_fetch_max(volatile atomic_intptr_t *object, uintptr_t opermax);
-intptr_t __ovld atomic_fetch_max_explicit(volatile atomic_intptr_t *object, uintptr_t opermax, memory_order minder);
-intptr_t __ovld atomic_fetch_max_explicit(volatile atomic_intptr_t *object, uintptr_t opermax, memory_order minder, memory_scope scope);
-#endif
-
-// atomic_store()
-
-void __ovld atomic_store(volatile atomic_int *object, int desired);
-void __ovld atomic_store_explicit(volatile atomic_int *object, int desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_int *object, int desired, memory_order order, memory_scope scope);
-void __ovld atomic_store(volatile atomic_uint *object, uint desired);
-void __ovld atomic_store_explicit(volatile atomic_uint *object, uint desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_uint *object, uint desired, memory_order order, memory_scope scope);
-void __ovld atomic_store(volatile atomic_float *object, float desired);
-void __ovld atomic_store_explicit(volatile atomic_float *object, float desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_float *object, float desired, memory_order order, memory_scope scope);
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#ifdef cl_khr_fp64
-void __ovld atomic_store(volatile atomic_double *object, double desired);
-void __ovld atomic_store_explicit(volatile atomic_double *object, double desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_double *object, double desired, memory_order order, memory_scope scope);
-#endif //cl_khr_fp64
-void __ovld atomic_store(volatile atomic_long *object, long desired);
-void __ovld atomic_store_explicit(volatile atomic_long *object, long desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_long *object, long desired, memory_order order, memory_scope scope);
-void __ovld atomic_store(volatile atomic_ulong *object, ulong desired);
-void __ovld atomic_store_explicit(volatile atomic_ulong *object, ulong desired, memory_order order);
-void __ovld atomic_store_explicit(volatile atomic_ulong *object, ulong desired, memory_order order, memory_scope scope);
-#endif
-
-// atomic_load()
-
-int __ovld atomic_load(volatile atomic_int *object);
-int __ovld atomic_load_explicit(volatile atomic_int *object, memory_order order);
-int __ovld atomic_load_explicit(volatile atomic_int *object, memory_order order, memory_scope scope);
-uint __ovld atomic_load(volatile atomic_uint *object);
-uint __ovld atomic_load_explicit(volatile atomic_uint *object, memory_order order);
-uint __ovld atomic_load_explicit(volatile atomic_uint *object, memory_order order, memory_scope scope);
-float __ovld atomic_load(volatile atomic_float *object);
-float __ovld atomic_load_explicit(volatile atomic_float *object, memory_order order);
-float __ovld atomic_load_explicit(volatile atomic_float *object, memory_order order, memory_scope scope);
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#ifdef cl_khr_fp64
-double __ovld atomic_load(volatile atomic_double *object);
-double __ovld atomic_load_explicit(volatile atomic_double *object, memory_order order);
-double __ovld atomic_load_explicit(volatile atomic_double *object, memory_order order, memory_scope scope);
-#endif //cl_khr_fp64
-long __ovld atomic_load(volatile atomic_long *object);
-long __ovld atomic_load_explicit(volatile atomic_long *object, memory_order order);
-long __ovld atomic_load_explicit(volatile atomic_long *object, memory_order order, memory_scope scope);
-ulong __ovld atomic_load(volatile atomic_ulong *object);
-ulong __ovld atomic_load_explicit(volatile atomic_ulong *object, memory_order order);
-ulong __ovld atomic_load_explicit(volatile atomic_ulong *object, memory_order order, memory_scope scope);
-#endif
-
-// atomic_exchange()
-
-int __ovld atomic_exchange(volatile atomic_int *object, int desired);
-int __ovld atomic_exchange_explicit(volatile atomic_int *object, int desired, memory_order order);
-int __ovld atomic_exchange_explicit(volatile atomic_int *object, int desired, memory_order order, memory_scope scope);
-uint __ovld atomic_exchange(volatile atomic_uint *object, uint desired);
-uint __ovld atomic_exchange_explicit(volatile atomic_uint *object, uint desired, memory_order order);
-uint __ovld atomic_exchange_explicit(volatile atomic_uint *object, uint desired, memory_order order, memory_scope scope);
-float __ovld atomic_exchange(volatile atomic_float *object, float desired);
-float __ovld atomic_exchange_explicit(volatile atomic_float *object, float desired, memory_order order);
-float __ovld atomic_exchange_explicit(volatile atomic_float *object, float desired, memory_order order, memory_scope scope);
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#ifdef cl_khr_fp64
-double __ovld atomic_exchange(volatile atomic_double *object, double desired);
-double __ovld atomic_exchange_explicit(volatile atomic_double *object, double desired, memory_order order);
-double __ovld atomic_exchange_explicit(volatile atomic_double *object, double desired, memory_order order, memory_scope scope);
-#endif //cl_khr_fp64
-long __ovld atomic_exchange(volatile atomic_long *object, long desired);
-long __ovld atomic_exchange_explicit(volatile atomic_long *object, long desired, memory_order order);
-long __ovld atomic_exchange_explicit(volatile atomic_long *object, long desired, memory_order order, memory_scope scope);
-ulong __ovld atomic_exchange(volatile atomic_ulong *object, ulong desired);
-ulong __ovld atomic_exchange_explicit(volatile atomic_ulong *object, ulong desired, memory_order order);
-ulong __ovld atomic_exchange_explicit(volatile atomic_ulong *object, ulong desired, memory_order order, memory_scope scope);
-#endif
-
-// atomic_compare_exchange_strong() and atomic_compare_exchange_weak()
-
-bool __ovld atomic_compare_exchange_strong(volatile atomic_int *object, int *expected, int desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_int *object, int *expected,
-                                                                                 int desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_int *object, int *expected,
-                                                                                 int desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_strong(volatile atomic_uint *object, uint *expected, uint desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_uint *object, uint *expected,
-                                                                                 uint desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_uint *object, uint *expected,
-                                                                                 uint desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_int *object, int *expected, int desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_int *object, int *expected,
-                                                                                 int desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_int *object, int *expected,
-                                                                                 int desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_uint *object, uint *expected, uint desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_uint *object, uint *expected,
-                                                                                 uint desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_uint *object, uint *expected,
-                                                                                 uint desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_strong(volatile atomic_float *object, float *expected, float desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_float *object, float *expected,
-                                                                                 float desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_float *object, float *expected,
-                                                                                 float desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_float *object, float *expected, float desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_float *object, float *expected,
-                                                                                 float desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_float *object, float *expected,
-                                                                                 float desired, memory_order success, memory_order failure, memory_scope scope);
-#if defined(cl_khr_int64_base_atomics) && defined(cl_khr_int64_extended_atomics)
-#ifdef cl_khr_fp64
-bool __ovld atomic_compare_exchange_strong(volatile atomic_double *object, double *expected, double desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_double *object, double *expected,
-                                                                                 double desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_double *object, double *expected,
-                                                                                 double desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_double *object, double *expected, double desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_double *object, double *expected,
-                                                                                 double desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_double *object, double *expected,
-                                                                                 double desired, memory_order success, memory_order failure, memory_scope scope);
-#endif //cl_khr_fp64
-bool __ovld atomic_compare_exchange_strong(volatile atomic_long *object, long *expected, long desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_long *object, long *expected,
-                                                                                 long desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_long *object, long *expected,
-                                                                                 long desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_long *object, long *expected, long desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_long *object, long *expected,
-                                                                                 long desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_long *object, long *expected,
-                                                                                 long desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_strong(volatile atomic_ulong *object, ulong *expected, ulong desired);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_ulong *object, ulong *expected,
-                                                                                 ulong desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_strong_explicit(volatile atomic_ulong *object, ulong *expected,
-                                                                                 ulong desired, memory_order success, memory_order failure, memory_scope scope);
-bool __ovld atomic_compare_exchange_weak(volatile atomic_ulong *object, ulong *expected, ulong desired);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_ulong *object, ulong *expected,
-                                                                                 ulong desired, memory_order success, memory_order failure);
-bool __ovld atomic_compare_exchange_weak_explicit(volatile atomic_ulong *object, ulong *expected,
-                                                                                 ulong desired, memory_order success, memory_order failure, memory_scope scope);
-#endif
-
-// atomic_flag_test_and_set() and atomic_flag_clear()
-
-bool __ovld atomic_flag_test_and_set(volatile atomic_flag *object);
-bool __ovld atomic_flag_test_and_set_explicit(volatile atomic_flag *object, memory_order order);
-bool __ovld atomic_flag_test_and_set_explicit(volatile atomic_flag *object, memory_order order, memory_scope scope);
-void __ovld atomic_flag_clear(volatile atomic_flag *object);
-void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order);
-void __ovld atomic_flag_clear_explicit(volatile atomic_flag *object, memory_order order, memory_scope scope);
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v1.1 s6.11.12, v1.2 s6.12.12, v2.0 s6.13.12 - Miscellaneous Vector Functions
-
-/**
- * The shuffle and shuffle2 built-in functions construct
- * a permutation of elements from one or two input
- * vectors respectively that are of the same type,
- * returning a vector with the same element type as the
- * input and length that is the same as the shuffle mask.
- * The size of each element in the mask must match the
- * size of each element in the result. For shuffle, only
- * the ilogb(2m-1) least significant bits of each mask
- * element are considered. For shuffle2, only the
- * ilogb(2m-1)+1 least significant bits of each mask
- * element are considered. Other bits in the mask shall
- * be ignored.
- * The elements of the input vectors are numbered from
- * left to right across one or both of the vectors. For this
- * purpose, the number of elements in a vector is given
- * by vec_step(gentypem). The shuffle mask operand
- * specifies, for each element of the result vector, which
- * element of the one or two input vectors the result
- * element gets.
- * Examples:
- * uint4 mask = (uint4)(3, 2,
- * 1, 0);
- * float4 a;
- * float4 r = shuffle(a, mask);
- * // r.s0123 = a.wzyx
- * uint8 mask = (uint8)(0, 1, 2, 3,
- * 4, 5, 6, 7);
- * float4 a, b;
- * float8 r = shuffle2(a, b, mask);
- * // r.s0123 = a.xyzw
- * // r.s4567 = b.xyzw
- * uint4 mask;
- * float8 a;
- * float4 b;
- * b = shuffle(a, mask);
- * Examples that are not valid are:
- * uint8 mask;
- * short16 a;
- * short8 b;
- * b = shuffle(a, mask); <- not valid
- */
-char2 __ovld __cnfn shuffle(char2 x, uchar2 mask);
-char2 __ovld __cnfn shuffle(char4 x, uchar2 mask);
-char2 __ovld __cnfn shuffle(char8 x, uchar2 mask);
-char2 __ovld __cnfn shuffle(char16 x, uchar2 mask);
-
-uchar2 __ovld __cnfn shuffle(uchar2 x, uchar2 mask);
-uchar2 __ovld __cnfn shuffle(uchar4 x, uchar2 mask);
-uchar2 __ovld __cnfn shuffle(uchar8 x, uchar2 mask);
-uchar2 __ovld __cnfn shuffle(uchar16 x, uchar2 mask);
-
-short2 __ovld __cnfn shuffle(short2 x, ushort2 mask);
-short2 __ovld __cnfn shuffle(short4 x, ushort2 mask);
-short2 __ovld __cnfn shuffle(short8 x, ushort2 mask);
-short2 __ovld __cnfn shuffle(short16 x, ushort2 mask);
-
-ushort2 __ovld __cnfn shuffle(ushort2 x, ushort2 mask);
-ushort2 __ovld __cnfn shuffle(ushort4 x, ushort2 mask);
-ushort2 __ovld __cnfn shuffle(ushort8 x, ushort2 mask);
-ushort2 __ovld __cnfn shuffle(ushort16 x, ushort2 mask);
-
-int2 __ovld __cnfn shuffle(int2 x, uint2 mask);
-int2 __ovld __cnfn shuffle(int4 x, uint2 mask);
-int2 __ovld __cnfn shuffle(int8 x, uint2 mask);
-int2 __ovld __cnfn shuffle(int16 x, uint2 mask);
-
-uint2 __ovld __cnfn shuffle(uint2 x, uint2 mask);
-uint2 __ovld __cnfn shuffle(uint4 x, uint2 mask);
-uint2 __ovld __cnfn shuffle(uint8 x, uint2 mask);
-uint2 __ovld __cnfn shuffle(uint16 x, uint2 mask);
-
-long2 __ovld __cnfn shuffle(long2 x, ulong2 mask);
-long2 __ovld __cnfn shuffle(long4 x, ulong2 mask);
-long2 __ovld __cnfn shuffle(long8 x, ulong2 mask);
-long2 __ovld __cnfn shuffle(long16 x, ulong2 mask);
-
-ulong2 __ovld __cnfn shuffle(ulong2 x, ulong2 mask);
-ulong2 __ovld __cnfn shuffle(ulong4 x, ulong2 mask);
-ulong2 __ovld __cnfn shuffle(ulong8 x, ulong2 mask);
-ulong2 __ovld __cnfn shuffle(ulong16 x, ulong2 mask);
-
-float2 __ovld __cnfn shuffle(float2 x, uint2 mask);
-float2 __ovld __cnfn shuffle(float4 x, uint2 mask);
-float2 __ovld __cnfn shuffle(float8 x, uint2 mask);
-float2 __ovld __cnfn shuffle(float16 x, uint2 mask);
-
-char4 __ovld __cnfn shuffle(char2 x, uchar4 mask);
-char4 __ovld __cnfn shuffle(char4 x, uchar4 mask);
-char4 __ovld __cnfn shuffle(char8 x, uchar4 mask);
-char4 __ovld __cnfn shuffle(char16 x, uchar4 mask);
-
-uchar4 __ovld __cnfn shuffle(uchar2 x, uchar4 mask);
-uchar4 __ovld __cnfn shuffle(uchar4 x, uchar4 mask);
-uchar4 __ovld __cnfn shuffle(uchar8 x, uchar4 mask);
-uchar4 __ovld __cnfn shuffle(uchar16 x, uchar4 mask);
-
-short4 __ovld __cnfn shuffle(short2 x, ushort4 mask);
-short4 __ovld __cnfn shuffle(short4 x, ushort4 mask);
-short4 __ovld __cnfn shuffle(short8 x, ushort4 mask);
-short4 __ovld __cnfn shuffle(short16 x, ushort4 mask);
-
-ushort4 __ovld __cnfn shuffle(ushort2 x, ushort4 mask);
-ushort4 __ovld __cnfn shuffle(ushort4 x, ushort4 mask);
-ushort4 __ovld __cnfn shuffle(ushort8 x, ushort4 mask);
-ushort4 __ovld __cnfn shuffle(ushort16 x, ushort4 mask);
-
-int4 __ovld __cnfn shuffle(int2 x, uint4 mask);
-int4 __ovld __cnfn shuffle(int4 x, uint4 mask);
-int4 __ovld __cnfn shuffle(int8 x, uint4 mask);
-int4 __ovld __cnfn shuffle(int16 x, uint4 mask);
-
-uint4 __ovld __cnfn shuffle(uint2 x, uint4 mask);
-uint4 __ovld __cnfn shuffle(uint4 x, uint4 mask);
-uint4 __ovld __cnfn shuffle(uint8 x, uint4 mask);
-uint4 __ovld __cnfn shuffle(uint16 x, uint4 mask);
-
-long4 __ovld __cnfn shuffle(long2 x, ulong4 mask);
-long4 __ovld __cnfn shuffle(long4 x, ulong4 mask);
-long4 __ovld __cnfn shuffle(long8 x, ulong4 mask);
-long4 __ovld __cnfn shuffle(long16 x, ulong4 mask);
-
-ulong4 __ovld __cnfn shuffle(ulong2 x, ulong4 mask);
-ulong4 __ovld __cnfn shuffle(ulong4 x, ulong4 mask);
-ulong4 __ovld __cnfn shuffle(ulong8 x, ulong4 mask);
-ulong4 __ovld __cnfn shuffle(ulong16 x, ulong4 mask);
-
-float4 __ovld __cnfn shuffle(float2 x, uint4 mask);
-float4 __ovld __cnfn shuffle(float4 x, uint4 mask);
-float4 __ovld __cnfn shuffle(float8 x, uint4 mask);
-float4 __ovld __cnfn shuffle(float16 x, uint4 mask);
-
-char8 __ovld __cnfn shuffle(char2 x, uchar8 mask);
-char8 __ovld __cnfn shuffle(char4 x, uchar8 mask);
-char8 __ovld __cnfn shuffle(char8 x, uchar8 mask);
-char8 __ovld __cnfn shuffle(char16 x, uchar8 mask);
-
-uchar8 __ovld __cnfn shuffle(uchar2 x, uchar8 mask);
-uchar8 __ovld __cnfn shuffle(uchar4 x, uchar8 mask);
-uchar8 __ovld __cnfn shuffle(uchar8 x, uchar8 mask);
-uchar8 __ovld __cnfn shuffle(uchar16 x, uchar8 mask);
-
-short8 __ovld __cnfn shuffle(short2 x, ushort8 mask);
-short8 __ovld __cnfn shuffle(short4 x, ushort8 mask);
-short8 __ovld __cnfn shuffle(short8 x, ushort8 mask);
-short8 __ovld __cnfn shuffle(short16 x, ushort8 mask);
-
-ushort8 __ovld __cnfn shuffle(ushort2 x, ushort8 mask);
-ushort8 __ovld __cnfn shuffle(ushort4 x, ushort8 mask);
-ushort8 __ovld __cnfn shuffle(ushort8 x, ushort8 mask);
-ushort8 __ovld __cnfn shuffle(ushort16 x, ushort8 mask);
-
-int8 __ovld __cnfn shuffle(int2 x, uint8 mask);
-int8 __ovld __cnfn shuffle(int4 x, uint8 mask);
-int8 __ovld __cnfn shuffle(int8 x, uint8 mask);
-int8 __ovld __cnfn shuffle(int16 x, uint8 mask);
-
-uint8 __ovld __cnfn shuffle(uint2 x, uint8 mask);
-uint8 __ovld __cnfn shuffle(uint4 x, uint8 mask);
-uint8 __ovld __cnfn shuffle(uint8 x, uint8 mask);
-uint8 __ovld __cnfn shuffle(uint16 x, uint8 mask);
-
-long8 __ovld __cnfn shuffle(long2 x, ulong8 mask);
-long8 __ovld __cnfn shuffle(long4 x, ulong8 mask);
-long8 __ovld __cnfn shuffle(long8 x, ulong8 mask);
-long8 __ovld __cnfn shuffle(long16 x, ulong8 mask);
-
-ulong8 __ovld __cnfn shuffle(ulong2 x, ulong8 mask);
-ulong8 __ovld __cnfn shuffle(ulong4 x, ulong8 mask);
-ulong8 __ovld __cnfn shuffle(ulong8 x, ulong8 mask);
-ulong8 __ovld __cnfn shuffle(ulong16 x, ulong8 mask);
-
-float8 __ovld __cnfn shuffle(float2 x, uint8 mask);
-float8 __ovld __cnfn shuffle(float4 x, uint8 mask);
-float8 __ovld __cnfn shuffle(float8 x, uint8 mask);
-float8 __ovld __cnfn shuffle(float16 x, uint8 mask);
-
-char16 __ovld __cnfn shuffle(char2 x, uchar16 mask);
-char16 __ovld __cnfn shuffle(char4 x, uchar16 mask);
-char16 __ovld __cnfn shuffle(char8 x, uchar16 mask);
-char16 __ovld __cnfn shuffle(char16 x, uchar16 mask);
-
-uchar16 __ovld __cnfn shuffle(uchar2 x, uchar16 mask);
-uchar16 __ovld __cnfn shuffle(uchar4 x, uchar16 mask);
-uchar16 __ovld __cnfn shuffle(uchar8 x, uchar16 mask);
-uchar16 __ovld __cnfn shuffle(uchar16 x, uchar16 mask);
-
-short16 __ovld __cnfn shuffle(short2 x, ushort16 mask);
-short16 __ovld __cnfn shuffle(short4 x, ushort16 mask);
-short16 __ovld __cnfn shuffle(short8 x, ushort16 mask);
-short16 __ovld __cnfn shuffle(short16 x, ushort16 mask);
-
-ushort16 __ovld __cnfn shuffle(ushort2 x, ushort16 mask);
-ushort16 __ovld __cnfn shuffle(ushort4 x, ushort16 mask);
-ushort16 __ovld __cnfn shuffle(ushort8 x, ushort16 mask);
-ushort16 __ovld __cnfn shuffle(ushort16 x, ushort16 mask);
-
-int16 __ovld __cnfn shuffle(int2 x, uint16 mask);
-int16 __ovld __cnfn shuffle(int4 x, uint16 mask);
-int16 __ovld __cnfn shuffle(int8 x, uint16 mask);
-int16 __ovld __cnfn shuffle(int16 x, uint16 mask);
-
-uint16 __ovld __cnfn shuffle(uint2 x, uint16 mask);
-uint16 __ovld __cnfn shuffle(uint4 x, uint16 mask);
-uint16 __ovld __cnfn shuffle(uint8 x, uint16 mask);
-uint16 __ovld __cnfn shuffle(uint16 x, uint16 mask);
-
-long16 __ovld __cnfn shuffle(long2 x, ulong16 mask);
-long16 __ovld __cnfn shuffle(long4 x, ulong16 mask);
-long16 __ovld __cnfn shuffle(long8 x, ulong16 mask);
-long16 __ovld __cnfn shuffle(long16 x, ulong16 mask);
-
-ulong16 __ovld __cnfn shuffle(ulong2 x, ulong16 mask);
-ulong16 __ovld __cnfn shuffle(ulong4 x, ulong16 mask);
-ulong16 __ovld __cnfn shuffle(ulong8 x, ulong16 mask);
-ulong16 __ovld __cnfn shuffle(ulong16 x, ulong16 mask);
-
-float16 __ovld __cnfn shuffle(float2 x, uint16 mask);
-float16 __ovld __cnfn shuffle(float4 x, uint16 mask);
-float16 __ovld __cnfn shuffle(float8 x, uint16 mask);
-float16 __ovld __cnfn shuffle(float16 x, uint16 mask);
-
-#ifdef cl_khr_fp64
-double2 __ovld __cnfn shuffle(double2 x, ulong2 mask);
-double2 __ovld __cnfn shuffle(double4 x, ulong2 mask);
-double2 __ovld __cnfn shuffle(double8 x, ulong2 mask);
-double2 __ovld __cnfn shuffle(double16 x, ulong2 mask);
-
-double4 __ovld __cnfn shuffle(double2 x, ulong4 mask);
-double4 __ovld __cnfn shuffle(double4 x, ulong4 mask);
-double4 __ovld __cnfn shuffle(double8 x, ulong4 mask);
-double4 __ovld __cnfn shuffle(double16 x, ulong4 mask);
-
-double8 __ovld __cnfn shuffle(double2 x, ulong8 mask);
-double8 __ovld __cnfn shuffle(double4 x, ulong8 mask);
-double8 __ovld __cnfn shuffle(double8 x, ulong8 mask);
-double8 __ovld __cnfn shuffle(double16 x, ulong8 mask);
-
-double16 __ovld __cnfn shuffle(double2 x, ulong16 mask);
-double16 __ovld __cnfn shuffle(double4 x, ulong16 mask);
-double16 __ovld __cnfn shuffle(double8 x, ulong16 mask);
-double16 __ovld __cnfn shuffle(double16 x, ulong16 mask);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half2 __ovld __cnfn shuffle(half2 x, ushort2 mask);
-half2 __ovld __cnfn shuffle(half4 x, ushort2 mask);
-half2 __ovld __cnfn shuffle(half8 x, ushort2 mask);
-half2 __ovld __cnfn shuffle(half16 x, ushort2 mask);
-
-half4 __ovld __cnfn shuffle(half2 x, ushort4 mask);
-half4 __ovld __cnfn shuffle(half4 x, ushort4 mask);
-half4 __ovld __cnfn shuffle(half8 x, ushort4 mask);
-half4 __ovld __cnfn shuffle(half16 x, ushort4 mask);
-
-half8 __ovld __cnfn shuffle(half2 x, ushort8 mask);
-half8 __ovld __cnfn shuffle(half4 x, ushort8 mask);
-half8 __ovld __cnfn shuffle(half8 x, ushort8 mask);
-half8 __ovld __cnfn shuffle(half16 x, ushort8 mask);
-
-half16 __ovld __cnfn shuffle(half2 x, ushort16 mask);
-half16 __ovld __cnfn shuffle(half4 x, ushort16 mask);
-half16 __ovld __cnfn shuffle(half8 x, ushort16 mask);
-half16 __ovld __cnfn shuffle(half16 x, ushort16 mask);
-#endif //cl_khr_fp16
-
-char2 __ovld __cnfn shuffle2(char2 x, char2 y, uchar2 mask);
-char2 __ovld __cnfn shuffle2(char4 x, char4 y, uchar2 mask);
-char2 __ovld __cnfn shuffle2(char8 x, char8 y, uchar2 mask);
-char2 __ovld __cnfn shuffle2(char16 x, char16 y, uchar2 mask);
-
-uchar2 __ovld __cnfn shuffle2(uchar2 x, uchar2 y, uchar2 mask);
-uchar2 __ovld __cnfn shuffle2(uchar4 x, uchar4 y, uchar2 mask);
-uchar2 __ovld __cnfn shuffle2(uchar8 x, uchar8 y, uchar2 mask);
-uchar2 __ovld __cnfn shuffle2(uchar16 x, uchar16 y, uchar2 mask);
-
-short2 __ovld __cnfn shuffle2(short2 x, short2 y, ushort2 mask);
-short2 __ovld __cnfn shuffle2(short4 x, short4 y, ushort2 mask);
-short2 __ovld __cnfn shuffle2(short8 x, short8 y, ushort2 mask);
-short2 __ovld __cnfn shuffle2(short16 x, short16 y, ushort2 mask);
-
-ushort2 __ovld __cnfn shuffle2(ushort2 x, ushort2 y, ushort2 mask);
-ushort2 __ovld __cnfn shuffle2(ushort4 x, ushort4 y, ushort2 mask);
-ushort2 __ovld __cnfn shuffle2(ushort8 x, ushort8 y, ushort2 mask);
-ushort2 __ovld __cnfn shuffle2(ushort16 x, ushort16 y, ushort2 mask);
-
-int2 __ovld __cnfn shuffle2(int2 x, int2 y, uint2 mask);
-int2 __ovld __cnfn shuffle2(int4 x, int4 y, uint2 mask);
-int2 __ovld __cnfn shuffle2(int8 x, int8 y, uint2 mask);
-int2 __ovld __cnfn shuffle2(int16 x, int16 y, uint2 mask);
-
-uint2 __ovld __cnfn shuffle2(uint2 x, uint2 y, uint2 mask);
-uint2 __ovld __cnfn shuffle2(uint4 x, uint4 y, uint2 mask);
-uint2 __ovld __cnfn shuffle2(uint8 x, uint8 y, uint2 mask);
-uint2 __ovld __cnfn shuffle2(uint16 x, uint16 y, uint2 mask);
-
-long2 __ovld __cnfn shuffle2(long2 x, long2 y, ulong2 mask);
-long2 __ovld __cnfn shuffle2(long4 x, long4 y, ulong2 mask);
-long2 __ovld __cnfn shuffle2(long8 x, long8 y, ulong2 mask);
-long2 __ovld __cnfn shuffle2(long16 x, long16 y, ulong2 mask);
-
-ulong2 __ovld __cnfn shuffle2(ulong2 x, ulong2 y, ulong2 mask);
-ulong2 __ovld __cnfn shuffle2(ulong4 x, ulong4 y, ulong2 mask);
-ulong2 __ovld __cnfn shuffle2(ulong8 x, ulong8 y, ulong2 mask);
-ulong2 __ovld __cnfn shuffle2(ulong16 x, ulong16 y, ulong2 mask);
-
-float2 __ovld __cnfn shuffle2(float2 x, float2 y, uint2 mask);
-float2 __ovld __cnfn shuffle2(float4 x, float4 y, uint2 mask);
-float2 __ovld __cnfn shuffle2(float8 x, float8 y, uint2 mask);
-float2 __ovld __cnfn shuffle2(float16 x, float16 y, uint2 mask);
-
-char4 __ovld __cnfn shuffle2(char2 x, char2 y, uchar4 mask);
-char4 __ovld __cnfn shuffle2(char4 x, char4 y, uchar4 mask);
-char4 __ovld __cnfn shuffle2(char8 x, char8 y, uchar4 mask);
-char4 __ovld __cnfn shuffle2(char16 x, char16 y, uchar4 mask);
-
-uchar4 __ovld __cnfn shuffle2(uchar2 x, uchar2 y, uchar4 mask);
-uchar4 __ovld __cnfn shuffle2(uchar4 x, uchar4 y, uchar4 mask);
-uchar4 __ovld __cnfn shuffle2(uchar8 x, uchar8 y, uchar4 mask);
-uchar4 __ovld __cnfn shuffle2(uchar16 x, uchar16 y, uchar4 mask);
-
-short4 __ovld __cnfn shuffle2(short2 x, short2 y, ushort4 mask);
-short4 __ovld __cnfn shuffle2(short4 x, short4 y, ushort4 mask);
-short4 __ovld __cnfn shuffle2(short8 x, short8 y, ushort4 mask);
-short4 __ovld __cnfn shuffle2(short16 x, short16 y, ushort4 mask);
-
-ushort4 __ovld __cnfn shuffle2(ushort2 x, ushort2 y, ushort4 mask);
-ushort4 __ovld __cnfn shuffle2(ushort4 x, ushort4 y, ushort4 mask);
-ushort4 __ovld __cnfn shuffle2(ushort8 x, ushort8 y, ushort4 mask);
-ushort4 __ovld __cnfn shuffle2(ushort16 x, ushort16 y, ushort4 mask);
-
-int4 __ovld __cnfn shuffle2(int2 x, int2 y, uint4 mask);
-int4 __ovld __cnfn shuffle2(int4 x, int4 y, uint4 mask);
-int4 __ovld __cnfn shuffle2(int8 x, int8 y, uint4 mask);
-int4 __ovld __cnfn shuffle2(int16 x, int16 y, uint4 mask);
-
-uint4 __ovld __cnfn shuffle2(uint2 x, uint2 y, uint4 mask);
-uint4 __ovld __cnfn shuffle2(uint4 x, uint4 y, uint4 mask);
-uint4 __ovld __cnfn shuffle2(uint8 x, uint8 y, uint4 mask);
-uint4 __ovld __cnfn shuffle2(uint16 x, uint16 y, uint4 mask);
-
-long4 __ovld __cnfn shuffle2(long2 x, long2 y, ulong4 mask);
-long4 __ovld __cnfn shuffle2(long4 x, long4 y, ulong4 mask);
-long4 __ovld __cnfn shuffle2(long8 x, long8 y, ulong4 mask);
-long4 __ovld __cnfn shuffle2(long16 x, long16 y, ulong4 mask);
-
-ulong4 __ovld __cnfn shuffle2(ulong2 x, ulong2 y, ulong4 mask);
-ulong4 __ovld __cnfn shuffle2(ulong4 x, ulong4 y, ulong4 mask);
-ulong4 __ovld __cnfn shuffle2(ulong8 x, ulong8 y, ulong4 mask);
-ulong4 __ovld __cnfn shuffle2(ulong16 x, ulong16 y, ulong4 mask);
-
-float4 __ovld __cnfn shuffle2(float2 x, float2 y, uint4 mask);
-float4 __ovld __cnfn shuffle2(float4 x, float4 y, uint4 mask);
-float4 __ovld __cnfn shuffle2(float8 x, float8 y, uint4 mask);
-float4 __ovld __cnfn shuffle2(float16 x, float16 y, uint4 mask);
-
-char8 __ovld __cnfn shuffle2(char2 x, char2 y, uchar8 mask);
-char8 __ovld __cnfn shuffle2(char4 x, char4 y, uchar8 mask);
-char8 __ovld __cnfn shuffle2(char8 x, char8 y, uchar8 mask);
-char8 __ovld __cnfn shuffle2(char16 x, char16 y, uchar8 mask);
-
-uchar8 __ovld __cnfn shuffle2(uchar2 x, uchar2 y, uchar8 mask);
-uchar8 __ovld __cnfn shuffle2(uchar4 x, uchar4 y, uchar8 mask);
-uchar8 __ovld __cnfn shuffle2(uchar8 x, uchar8 y, uchar8 mask);
-uchar8 __ovld __cnfn shuffle2(uchar16 x, uchar16 y, uchar8 mask);
-
-short8 __ovld __cnfn shuffle2(short2 x, short2 y, ushort8 mask);
-short8 __ovld __cnfn shuffle2(short4 x, short4 y, ushort8 mask);
-short8 __ovld __cnfn shuffle2(short8 x, short8 y, ushort8 mask);
-short8 __ovld __cnfn shuffle2(short16 x, short16 y, ushort8 mask);
-
-ushort8 __ovld __cnfn shuffle2(ushort2 x, ushort2 y, ushort8 mask);
-ushort8 __ovld __cnfn shuffle2(ushort4 x, ushort4 y, ushort8 mask);
-ushort8 __ovld __cnfn shuffle2(ushort8 x, ushort8 y, ushort8 mask);
-ushort8 __ovld __cnfn shuffle2(ushort16 x, ushort16 y, ushort8 mask);
-
-int8 __ovld __cnfn shuffle2(int2 x, int2 y, uint8 mask);
-int8 __ovld __cnfn shuffle2(int4 x, int4 y, uint8 mask);
-int8 __ovld __cnfn shuffle2(int8 x, int8 y, uint8 mask);
-int8 __ovld __cnfn shuffle2(int16 x, int16 y, uint8 mask);
-
-uint8 __ovld __cnfn shuffle2(uint2 x, uint2 y, uint8 mask);
-uint8 __ovld __cnfn shuffle2(uint4 x, uint4 y, uint8 mask);
-uint8 __ovld __cnfn shuffle2(uint8 x, uint8 y, uint8 mask);
-uint8 __ovld __cnfn shuffle2(uint16 x, uint16 y, uint8 mask);
-
-long8 __ovld __cnfn shuffle2(long2 x, long2 y, ulong8 mask);
-long8 __ovld __cnfn shuffle2(long4 x, long4 y, ulong8 mask);
-long8 __ovld __cnfn shuffle2(long8 x, long8 y, ulong8 mask);
-long8 __ovld __cnfn shuffle2(long16 x, long16 y, ulong8 mask);
-
-ulong8 __ovld __cnfn shuffle2(ulong2 x, ulong2 y, ulong8 mask);
-ulong8 __ovld __cnfn shuffle2(ulong4 x, ulong4 y, ulong8 mask);
-ulong8 __ovld __cnfn shuffle2(ulong8 x, ulong8 y, ulong8 mask);
-ulong8 __ovld __cnfn shuffle2(ulong16 x, ulong16 y, ulong8 mask);
-
-float8 __ovld __cnfn shuffle2(float2 x, float2 y, uint8 mask);
-float8 __ovld __cnfn shuffle2(float4 x, float4 y, uint8 mask);
-float8 __ovld __cnfn shuffle2(float8 x, float8 y, uint8 mask);
-float8 __ovld __cnfn shuffle2(float16 x, float16 y, uint8 mask);
-
-char16 __ovld __cnfn shuffle2(char2 x, char2 y, uchar16 mask);
-char16 __ovld __cnfn shuffle2(char4 x, char4 y, uchar16 mask);
-char16 __ovld __cnfn shuffle2(char8 x, char8 y, uchar16 mask);
-char16 __ovld __cnfn shuffle2(char16 x, char16 y, uchar16 mask);
-
-uchar16 __ovld __cnfn shuffle2(uchar2 x, uchar2 y, uchar16 mask);
-uchar16 __ovld __cnfn shuffle2(uchar4 x, uchar4 y, uchar16 mask);
-uchar16 __ovld __cnfn shuffle2(uchar8 x, uchar8 y, uchar16 mask);
-uchar16 __ovld __cnfn shuffle2(uchar16 x, uchar16 y, uchar16 mask);
-
-short16 __ovld __cnfn shuffle2(short2 x, short2 y, ushort16 mask);
-short16 __ovld __cnfn shuffle2(short4 x, short4 y, ushort16 mask);
-short16 __ovld __cnfn shuffle2(short8 x, short8 y, ushort16 mask);
-short16 __ovld __cnfn shuffle2(short16 x, short16 y, ushort16 mask);
-
-ushort16 __ovld __cnfn shuffle2(ushort2 x, ushort2 y, ushort16 mask);
-ushort16 __ovld __cnfn shuffle2(ushort4 x, ushort4 y, ushort16 mask);
-ushort16 __ovld __cnfn shuffle2(ushort8 x, ushort8 y, ushort16 mask);
-ushort16 __ovld __cnfn shuffle2(ushort16 x, ushort16 y, ushort16 mask);
-
-int16 __ovld __cnfn shuffle2(int2 x, int2 y, uint16 mask);
-int16 __ovld __cnfn shuffle2(int4 x, int4 y, uint16 mask);
-int16 __ovld __cnfn shuffle2(int8 x, int8 y, uint16 mask);
-int16 __ovld __cnfn shuffle2(int16 x, int16 y, uint16 mask);
-
-uint16 __ovld __cnfn shuffle2(uint2 x, uint2 y, uint16 mask);
-uint16 __ovld __cnfn shuffle2(uint4 x, uint4 y, uint16 mask);
-uint16 __ovld __cnfn shuffle2(uint8 x, uint8 y, uint16 mask);
-uint16 __ovld __cnfn shuffle2(uint16 x, uint16 y, uint16 mask);
-
-long16 __ovld __cnfn shuffle2(long2 x, long2 y, ulong16 mask);
-long16 __ovld __cnfn shuffle2(long4 x, long4 y, ulong16 mask);
-long16 __ovld __cnfn shuffle2(long8 x, long8 y, ulong16 mask);
-long16 __ovld __cnfn shuffle2(long16 x, long16 y, ulong16 mask);
-
-ulong16 __ovld __cnfn shuffle2(ulong2 x, ulong2 y, ulong16 mask);
-ulong16 __ovld __cnfn shuffle2(ulong4 x, ulong4 y, ulong16 mask);
-ulong16 __ovld __cnfn shuffle2(ulong8 x, ulong8 y, ulong16 mask);
-ulong16 __ovld __cnfn shuffle2(ulong16 x, ulong16 y, ulong16 mask);
-
-float16 __ovld __cnfn shuffle2(float2 x, float2 y, uint16 mask);
-float16 __ovld __cnfn shuffle2(float4 x, float4 y, uint16 mask);
-float16 __ovld __cnfn shuffle2(float8 x, float8 y, uint16 mask);
-float16 __ovld __cnfn shuffle2(float16 x, float16 y, uint16 mask);
-
-#ifdef cl_khr_fp64
-double2 __ovld __cnfn shuffle2(double2 x, double2 y, ulong2 mask);
-double2 __ovld __cnfn shuffle2(double4 x, double4 y, ulong2 mask);
-double2 __ovld __cnfn shuffle2(double8 x, double8 y, ulong2 mask);
-double2 __ovld __cnfn shuffle2(double16 x, double16 y, ulong2 mask);
-
-double4 __ovld __cnfn shuffle2(double2 x, double2 y, ulong4 mask);
-double4 __ovld __cnfn shuffle2(double4 x, double4 y, ulong4 mask);
-double4 __ovld __cnfn shuffle2(double8 x, double8 y, ulong4 mask);
-double4 __ovld __cnfn shuffle2(double16 x, double16 y, ulong4 mask);
-
-double8 __ovld __cnfn shuffle2(double2 x, double2 y, ulong8 mask);
-double8 __ovld __cnfn shuffle2(double4 x, double4 y, ulong8 mask);
-double8 __ovld __cnfn shuffle2(double8 x, double8 y, ulong8 mask);
-double8 __ovld __cnfn shuffle2(double16 x, double16 y, ulong8 mask);
-
-double16 __ovld __cnfn shuffle2(double2 x, double2 y, ulong16 mask);
-double16 __ovld __cnfn shuffle2(double4 x, double4 y, ulong16 mask);
-double16 __ovld __cnfn shuffle2(double8 x, double8 y, ulong16 mask);
-double16 __ovld __cnfn shuffle2(double16 x, double16 y, ulong16 mask);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half2 __ovld __cnfn shuffle2(half2 x, half2 y, ushort2 mask);
-half2 __ovld __cnfn shuffle2(half4 x, half4 y, ushort2 mask);
-half2 __ovld __cnfn shuffle2(half8 x, half8 y, ushort2 mask);
-half2 __ovld __cnfn shuffle2(half16 x, half16 y, ushort2 mask);
-
-half4 __ovld __cnfn shuffle2(half2 x, half2 y, ushort4 mask);
-half4 __ovld __cnfn shuffle2(half4 x, half4 y, ushort4 mask);
-half4 __ovld __cnfn shuffle2(half8 x, half8 y, ushort4 mask);
-half4 __ovld __cnfn shuffle2(half16 x, half16 y, ushort4 mask);
-
-half8 __ovld __cnfn shuffle2(half2 x, half2 y, ushort8 mask);
-half8 __ovld __cnfn shuffle2(half4 x, half4 y, ushort8 mask);
-half8 __ovld __cnfn shuffle2(half8 x, half8 y, ushort8 mask);
-half8 __ovld __cnfn shuffle2(half16 x, half16 y, ushort8 mask);
-
-half16 __ovld __cnfn shuffle2(half2 x, half2 y, ushort16 mask);
-half16 __ovld __cnfn shuffle2(half4 x, half4 y, ushort16 mask);
-half16 __ovld __cnfn shuffle2(half8 x, half8 y, ushort16 mask);
-half16 __ovld __cnfn shuffle2(half16 x, half16 y, ushort16 mask);
-#endif //cl_khr_fp16
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-// OpenCL v1.2 s6.12.13, v2.0 s6.13.13 - printf
-
-int printf(__constant const char* st, ...) __attribute__((format(printf, 1, 2)));
-#endif
-
-// OpenCL v1.1 s6.11.3, v1.2 s6.12.14, v2.0 s6.13.14 - Image Read and Write Functions
-
-#ifdef cl_khr_gl_msaa_sharing
-#pragma OPENCL EXTENSION cl_khr_gl_msaa_sharing : enable
-#endif //cl_khr_gl_msaa_sharing
-
-/**
- * Use the coordinate (coord.xy) to do an element lookup in
- * the 2D image object specified by image.
- *
- * Use the coordinate (coord.x, coord.y, coord.z) to do
- * an element lookup in the 3D image object specified
- * by image. coord.w is ignored.
- *
- * Use the coordinate (coord.z) to index into the
- * 2D image array object specified by image_array
- * and (coord.x, coord.y) to do an element lookup in
- * the 2D image object specified by image.
- *
- * Use the coordinate (x) to do an element lookup in
- * the 1D image object specified by image.
- *
- * Use the coordinate (coord.y) to index into the
- * 1D image array object specified by image_array
- * and (coord.x) to do an element lookup in
- * the 1D image object specified by image.
- *
- * Use the coordinate (cood.xy) and sample to do an
- * element lookup in the 2D multi-sample image specified
- * by image.
- *
- * Use coord.xy and sample to do an element
- * lookup in the 2D multi-sample image layer
- * identified by index coord.z in the 2D multi-sample
- * image array specified by image.
- *
- * For mipmap images, use the mip-level specified by
- * the Level-of-Detail (lod) or use gradients for LOD
- * computation.
- *
- * read_imagef returns floating-point values in the
- * range [0.0 ... 1.0] for image objects created with
- * image_channel_data_type set to one of the predefined
- * packed formats or CL_UNORM_INT8, or
- * CL_UNORM_INT16.
- *
- * read_imagef returns floating-point values in the
- * range [-1.0 ... 1.0] for image objects created with
- * image_channel_data_type set to CL_SNORM_INT8,
- * or CL_SNORM_INT16.
- *
- * read_imagef returns floating-point values for image
- * objects created with image_channel_data_type set to
- * CL_HALF_FLOAT or CL_FLOAT.
- *
- * read_imagei and read_imageui return
- * unnormalized signed integer and unsigned integer
- * values respectively. Each channel will be stored in a
- * 32-bit integer.
- *
- * read_imagei can only be used with image objects
- * created with image_channel_data_type set to one of
- * the following values:
- * CL_SIGNED_INT8,
- * CL_SIGNED_INT16 and
- * CL_SIGNED_INT32.
- * If the image_channel_data_type is not one of the
- * above values, the values returned by read_imagei
- * are undefined.
- *
- * read_imageui can only be used with image objects
- * created with image_channel_data_type set to one of
- * the following values:
- * CL_UNSIGNED_INT8,
- * CL_UNSIGNED_INT16 and
- * CL_UNSIGNED_INT32.
- * If the image_channel_data_type is not one of the
- * above values, the values returned by read_imageui
- * are undefined.
- *
- * The read_image{i|ui} calls support a nearest filter
- * only. The filter_mode specified in sampler
- * must be set to CLK_FILTER_NEAREST; otherwise
- * the values returned are undefined.
-
- * The read_image{f|i|ui} calls that take
- * integer coordinates must use a sampler with
- * normalized coordinates set to
- * CLK_NORMALIZED_COORDS_FALSE and
- * addressing mode set to
- * CLK_ADDRESS_CLAMP_TO_EDGE,
- * CLK_ADDRESS_CLAMP or CLK_ADDRESS_NONE;
- * otherwise the values returned are undefined.
- *
- * Values returned by read_imagef for image objects
- * with image_channel_data_type values not specified
- * in the description above are undefined.
- */
-
-float4 __purefn __ovld read_imagef(read_only image2d_t image, sampler_t sampler, int2 coord);
-float4 __purefn __ovld read_imagef(read_only image2d_t image, sampler_t sampler, float2 coord);
-
-int4 __purefn __ovld read_imagei(read_only image2d_t image, sampler_t sampler, int2 coord);
-int4 __purefn __ovld read_imagei(read_only image2d_t image, sampler_t sampler, float2 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_t image, sampler_t sampler, int2 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_t image, sampler_t sampler, float2 coord);
-
-float4 __purefn __ovld read_imagef(read_only image3d_t image, sampler_t sampler, int4 coord);
-float4 __purefn __ovld read_imagef(read_only image3d_t image, sampler_t sampler, float4 coord);
-
-int4 __purefn __ovld read_imagei(read_only image3d_t image, sampler_t sampler, int4 coord);
-int4 __purefn __ovld read_imagei(read_only image3d_t image, sampler_t sampler, float4 coord);
-uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, int4 coord);
-uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord);
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
-float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
-
-int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
-int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, int4 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, float4 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-
-float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, int coord);
-float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord);
-
-int4 __purefn __ovld read_imagei(read_only image1d_t image, sampler_t sampler, int coord);
-int4 __purefn __ovld read_imagei(read_only image1d_t image, sampler_t sampler, float coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, int coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, float coord);
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
-float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
-
-int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
-int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, int2 coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, float2 coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-
-#ifdef cl_khr_depth_images
-float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, float2 coord);
-float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, int2 coord);
-
-float __purefn __ovld read_imagef(read_only image2d_array_depth_t image, sampler_t sampler, float4 coord);
-float __purefn __ovld read_imagef(read_only image2d_array_depth_t image, sampler_t sampler, int4 coord);
-#endif //cl_khr_depth_images
-
-#if defined(cl_khr_gl_msaa_sharing)
-float4 __purefn __ovld read_imagef(read_only image2d_msaa_t image, int2 coord, int sample);
-int4 __purefn __ovld read_imagei(read_only image2d_msaa_t image, int2 coord, int sample);
-uint4 __purefn __ovld read_imageui(read_only image2d_msaa_t image, int2 coord, int sample);
-
-float __purefn __ovld read_imagef(read_only image2d_msaa_depth_t image, int2 coord, int sample);
-
-float4 __purefn __ovld read_imagef(read_only image2d_array_msaa_t image, int4 coord, int sample);
-int4 __purefn __ovld read_imagei(read_only image2d_array_msaa_t image, int4 coord, int sample);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_msaa_t image, int4 coord, int sample);
-
-float __purefn __ovld read_imagef(read_only image2d_array_msaa_depth_t image, int4 coord, int sample);
-#endif //cl_khr_gl_msaa_sharing
-
-// OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifdef cl_khr_mipmap_image
-
-float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord, float lod);
-int4 __purefn __ovld read_imagei(read_only image1d_t image, sampler_t sampler, float coord, float lod);
-uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, float coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_only image2d_t image, sampler_t sampler, float2 coord, float lod);
-int4 __purefn __ovld read_imagei(read_only image2d_t image, sampler_t sampler, float2 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_only image2d_t image, sampler_t sampler, float2 coord, float lod);
-
-float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, float2 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-
-float __purefn __ovld read_imagef(read_only image2d_array_depth_t image, sampler_t sampler, float4 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_only image3d_t image, sampler_t sampler, float4 coord, float lod);
-int4 __purefn __ovld read_imagei(read_only image3d_t image, sampler_t sampler, float4 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_only image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-int4 __purefn __ovld read_imagei(read_only image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-uint4 __purefn __ovld read_imageui(read_only image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-
-float4 __purefn __ovld read_imagef(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-int4 __purefn __ovld read_imagei(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-uint4 __purefn __ovld read_imageui(read_only image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-
-float4 __purefn __ovld read_imagef(read_only image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-int4 __purefn __ovld read_imagei(read_only image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-uint4 __purefn __ovld read_imageui(read_only image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-
-float __purefn __ovld read_imagef(read_only image2d_depth_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-
-float4 __purefn __ovld read_imagef(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-int4 __purefn __ovld read_imagei(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-
-float __purefn __ovld read_imagef(read_only image2d_array_depth_t image, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-
-float4 __purefn __ovld read_imagef(read_only image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-int4 __purefn __ovld read_imagei(read_only image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-uint4 __purefn __ovld read_imageui(read_only image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-
-#endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-
-/**
-* Sampler-less Image Access
-*/
-
-float4 __purefn __ovld read_imagef(read_only image1d_t image, int coord);
-int4 __purefn __ovld read_imagei(read_only image1d_t image, int coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_t image, int coord);
-
-float4 __purefn __ovld read_imagef(read_only image1d_buffer_t image, int coord);
-int4 __purefn __ovld read_imagei(read_only image1d_buffer_t image, int coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_buffer_t image, int coord);
-
-float4 __purefn __ovld read_imagef(read_only image1d_array_t image, int2 coord);
-int4 __purefn __ovld read_imagei(read_only image1d_array_t image, int2 coord);
-uint4 __purefn __ovld read_imageui(read_only image1d_array_t image, int2 coord);
-
-float4 __purefn __ovld read_imagef(read_only image2d_t image, int2 coord);
-int4 __purefn __ovld read_imagei(read_only image2d_t image, int2 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_t image, int2 coord);
-
-float4 __purefn __ovld read_imagef(read_only image2d_array_t image, int4 coord);
-int4 __purefn __ovld read_imagei(read_only image2d_array_t image, int4 coord);
-uint4 __purefn __ovld read_imageui(read_only image2d_array_t image, int4 coord);
-
-#ifdef cl_khr_depth_images
-float __purefn __ovld read_imagef(read_only image2d_depth_t image, int2 coord);
-float __purefn __ovld read_imagef(read_only image2d_array_depth_t image, int4 coord);
-#endif //cl_khr_depth_images
-
-float4 __purefn __ovld read_imagef(read_only image3d_t image, int4 coord);
-int4 __purefn __ovld read_imagei(read_only image3d_t image, int4 coord);
-uint4 __purefn __ovld read_imageui(read_only image3d_t image, int4 coord);
-
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-
-// Image read functions returning half4 type
-#ifdef cl_khr_fp16
-half4 __purefn __ovld read_imageh(read_only image1d_t image, sampler_t sampler, int coord);
-half4 __purefn __ovld read_imageh(read_only image1d_t image, sampler_t sampler, float coord);
-half4 __purefn __ovld read_imageh(read_only image2d_t image, sampler_t sampler, int2 coord);
-half4 __purefn __ovld read_imageh(read_only image2d_t image, sampler_t sampler, float2 coord);
-half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, int4 coord);
-half4 __purefn __ovld read_imageh(read_only image3d_t image, sampler_t sampler, float4 coord);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, int2 coord);
-half4 __purefn __ovld read_imageh(read_only image1d_array_t image, sampler_t sampler, float2 coord);
-half4 __purefn __ovld read_imageh(read_only image2d_array_t image, sampler_t sampler, int4 coord);
-half4 __purefn __ovld read_imageh(read_only image2d_array_t image, sampler_t sampler, float4 coord);
-/**
- * Sampler-less Image Access
- */
-half4 __purefn __ovld read_imageh(read_only image1d_t image, int coord);
-half4 __purefn __ovld read_imageh(read_only image2d_t image, int2 coord);
-half4 __purefn __ovld read_imageh(read_only image3d_t image, int4 coord);
-half4 __purefn __ovld read_imageh(read_only image1d_array_t image, int2 coord);
-half4 __purefn __ovld read_imageh(read_only image2d_array_t image, int4 coord);
-half4 __purefn __ovld read_imageh(read_only image1d_buffer_t image, int coord);
-#endif // __OPENCL_C_VERSION__ >= CL_VERSION_1_2
-#endif //cl_khr_fp16
-
-// Image read functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-float4 __purefn __ovld read_imagef(read_write image1d_t image, int coord);
-int4 __purefn __ovld read_imagei(read_write image1d_t image, int coord);
-uint4 __purefn __ovld read_imageui(read_write image1d_t image, int coord);
-
-float4 __purefn __ovld read_imagef(read_write image1d_buffer_t image, int coord);
-int4 __purefn __ovld read_imagei(read_write image1d_buffer_t image, int coord);
-uint4 __purefn __ovld read_imageui(read_write image1d_buffer_t image, int coord);
-
-float4 __purefn __ovld read_imagef(read_write image1d_array_t image, int2 coord);
-int4 __purefn __ovld read_imagei(read_write image1d_array_t image, int2 coord);
-uint4 __purefn __ovld read_imageui(read_write image1d_array_t image, int2 coord);
-
-float4 __purefn __ovld read_imagef(read_write image2d_t image, int2 coord);
-int4 __purefn __ovld read_imagei(read_write image2d_t image, int2 coord);
-uint4 __purefn __ovld read_imageui(read_write image2d_t image, int2 coord);
-
-float4 __purefn __ovld read_imagef(read_write image2d_array_t image, int4 coord);
-int4 __purefn __ovld read_imagei(read_write image2d_array_t image, int4 coord);
-uint4 __purefn __ovld read_imageui(read_write image2d_array_t image, int4 coord);
-
-float4 __purefn __ovld read_imagef(read_write image3d_t image, int4 coord);
-int4 __purefn __ovld read_imagei(read_write image3d_t image, int4 coord);
-uint4 __purefn __ovld read_imageui(read_write image3d_t image, int4 coord);
-
-#ifdef cl_khr_depth_images
-float __purefn __ovld read_imagef(read_write image2d_depth_t image, int2 coord);
-float __purefn __ovld read_imagef(read_write image2d_array_depth_t image, int4 coord);
-#endif //cl_khr_depth_images
-
-#if cl_khr_gl_msaa_sharing
-float4 __purefn __ovld read_imagef(read_write image2d_msaa_t image, int2 coord, int sample);
-int4 __purefn __ovld read_imagei(read_write image2d_msaa_t image, int2 coord, int sample);
-uint4 __purefn __ovld read_imageui(read_write image2d_msaa_t image, int2 coord, int sample);
-
-float4 __purefn __ovld read_imagef(read_write image2d_array_msaa_t image, int4 coord, int sample);
-int4 __purefn __ovld read_imagei(read_write image2d_array_msaa_t image, int4 coord, int sample);
-uint4 __purefn __ovld read_imageui(read_write image2d_array_msaa_t image, int4 coord, int sample);
-
-float __purefn __ovld read_imagef(read_write image2d_msaa_depth_t image, int2 coord, int sample);
-float __purefn __ovld read_imagef(read_write image2d_array_msaa_depth_t image, int4 coord, int sample);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifdef cl_khr_mipmap_image
-float4 __purefn __ovld read_imagef(read_write image1d_t image, sampler_t sampler, float coord, float lod);
-int4 __purefn __ovld read_imagei(read_write image1d_t image, sampler_t sampler, float coord, float lod);
-uint4 __purefn __ovld read_imageui(read_write image1d_t image, sampler_t sampler, float coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-int4 __purefn __ovld read_imagei(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_write image2d_t image, sampler_t sampler, float2 coord, float lod);
-int4 __purefn __ovld read_imagei(read_write image2d_t image, sampler_t sampler, float2 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_write image2d_t image, sampler_t sampler, float2 coord, float lod);
-
-float __purefn __ovld read_imagef(read_write image2d_depth_t image, sampler_t sampler, float2 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-int4 __purefn __ovld read_imagei(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float lod);
-
-float __purefn __ovld read_imagef(read_write image2d_array_depth_t image, sampler_t sampler, float4 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_write image3d_t image, sampler_t sampler, float4 coord, float lod);
-int4 __purefn __ovld read_imagei(read_write image3d_t image, sampler_t sampler, float4 coord, float lod);
-uint4 __purefn __ovld read_imageui(read_write image3d_t image, sampler_t sampler, float4 coord, float lod);
-
-float4 __purefn __ovld read_imagef(read_write image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-int4 __purefn __ovld read_imagei(read_write image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-uint4 __purefn __ovld read_imageui(read_write image1d_t image, sampler_t sampler, float coord, float gradientX, float gradientY);
-
-float4 __purefn __ovld read_imagef(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-int4 __purefn __ovld read_imagei(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-uint4 __purefn __ovld read_imageui(read_write image1d_array_t image_array, sampler_t sampler, float2 coord, float gradientX, float gradientY);
-
-float4 __purefn __ovld read_imagef(read_write image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-int4 __purefn __ovld read_imagei(read_write image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-uint4 __purefn __ovld read_imageui(read_write image2d_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-
-float __purefn __ovld read_imagef(read_write image2d_depth_t image, sampler_t sampler, float2 coord, float2 gradientX, float2 gradientY);
-
-float4 __purefn __ovld read_imagef(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-int4 __purefn __ovld read_imagei(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-uint4 __purefn __ovld read_imageui(read_write image2d_array_t image_array, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-
-float __purefn __ovld read_imagef(read_write image2d_array_depth_t image, sampler_t sampler, float4 coord, float2 gradientX, float2 gradientY);
-
-float4 __purefn __ovld read_imagef(read_write image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-int4 __purefn __ovld read_imagei(read_write image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-uint4 __purefn __ovld read_imageui(read_write image3d_t image, sampler_t sampler, float4 coord, float4 gradientX, float4 gradientY);
-
-#endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// Image read functions returning half4 type
-#ifdef cl_khr_fp16
-half4 __purefn __ovld read_imageh(read_write image1d_t image, int coord);
-half4 __purefn __ovld read_imageh(read_write image2d_t image, int2 coord);
-half4 __purefn __ovld read_imageh(read_write image3d_t image, int4 coord);
-half4 __purefn __ovld read_imageh(read_write image1d_array_t image, int2 coord);
-half4 __purefn __ovld read_imageh(read_write image2d_array_t image, int4 coord);
-half4 __purefn __ovld read_imageh(read_write image1d_buffer_t image, int coord);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Write color value to location specified by coordinate
- * (coord.x, coord.y) in the 2D image object specified by image.
- * (coord.x, coord.y) are considered to be unnormalized coordinates
- * and must be in the range 0 ... image width - 1, and 0
- * ... image height - 1.
-
- * Write color value to location specified by coordinate
- * (coord.x, coord.y) in the 2D image object specified by index
- * (coord.z) of the 2D image array object image_array.
- * (coord.x, coord.y) are considered to be unnormalized
- * coordinates and must be in the range 0 ... image width
- * - 1.
- *
- * Write color value to location specified by coordinate
- * (coord) in the 1D image (buffer) object specified by image.
- * coord is considered to be unnormalized coordinates
- * and must be in the range 0 ... image width - 1.
- *
- * Write color value to location specified by coordinate
- * (coord.x) in the 1D image object specified by index
- * (coord.y) of the 1D image array object image_array.
- * x is considered to be unnormalized coordinates
- * and must be in the range 0 ... image width - 1.
- *
- * Write color value to location specified by coordinate
- * (coord.x, coord.y, coord.z) in the 3D image object specified by image.
- * coord.x & coord.y are considered to be unnormalized coordinates
- * and must be in the range 0 ... image width - 1, and 0
- * ... image height - 1.
- *
- * For mipmap images, use mip-level specified by lod.
- *
- * Appropriate data format conversion to the specified
- * image format is done before writing the color value.
- *
- * write_imagef can only be used with image objects
- * created with image_channel_data_type set to one of
- * the pre-defined packed formats or set to
- * CL_SNORM_INT8, CL_UNORM_INT8,
- * CL_SNORM_INT16, CL_UNORM_INT16,
- * CL_HALF_FLOAT or CL_FLOAT. Appropriate data
- * format conversion will be done to convert channel
- * data from a floating-point value to actual data format
- * in which the channels are stored.
- *
- * write_imagei can only be used with image objects
- * created with image_channel_data_type set to one of
- * the following values:
- * CL_SIGNED_INT8,
- * CL_SIGNED_INT16 and
- * CL_SIGNED_INT32.
- *
- * write_imageui can only be used with image objects
- * created with image_channel_data_type set to one of
- * the following values:
- * CL_UNSIGNED_INT8,
- * CL_UNSIGNED_INT16 and
- * CL_UNSIGNED_INT32.
- *
- * The behavior of write_imagef, write_imagei and
- * write_imageui for image objects created with
- * image_channel_data_type values not specified in
- * the description above or with (x, y) coordinate
- * values that are not in the range (0 ... image width -1,
- * 0 ... image height - 1), respectively, is undefined.
- */
-void __ovld write_imagef(write_only image2d_t image, int2 coord, float4 color);
-void __ovld write_imagei(write_only image2d_t image, int2 coord, int4 color);
-void __ovld write_imageui(write_only image2d_t image, int2 coord, uint4 color);
-
-void __ovld write_imagef(write_only image2d_array_t image_array, int4 coord, float4 color);
-void __ovld write_imagei(write_only image2d_array_t image_array, int4 coord, int4 color);
-void __ovld write_imageui(write_only image2d_array_t image_array, int4 coord, uint4 color);
-
-void __ovld write_imagef(write_only image1d_t image, int coord, float4 color);
-void __ovld write_imagei(write_only image1d_t image, int coord, int4 color);
-void __ovld write_imageui(write_only image1d_t image, int coord, uint4 color);
-
-void __ovld write_imagef(write_only image1d_buffer_t image, int coord, float4 color);
-void __ovld write_imagei(write_only image1d_buffer_t image, int coord, int4 color);
-void __ovld write_imageui(write_only image1d_buffer_t image, int coord, uint4 color);
-
-void __ovld write_imagef(write_only image1d_array_t image_array, int2 coord, float4 color);
-void __ovld write_imagei(write_only image1d_array_t image_array, int2 coord, int4 color);
-void __ovld write_imageui(write_only image1d_array_t image_array, int2 coord, uint4 color);
-
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imagef(write_only image3d_t image, int4 coord, float4 color);
-void __ovld write_imagei(write_only image3d_t image, int4 coord, int4 color);
-void __ovld write_imageui(write_only image3d_t image, int4 coord, uint4 color);
-#endif
-
-#ifdef cl_khr_depth_images
-void __ovld write_imagef(write_only image2d_depth_t image, int2 coord, float color);
-void __ovld write_imagef(write_only image2d_array_depth_t image, int4 coord, float color);
-#endif //cl_khr_depth_images
-
-// OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifdef cl_khr_mipmap_image
-void __ovld write_imagef(write_only image1d_t image, int coord, int lod, float4 color);
-void __ovld write_imagei(write_only image1d_t image, int coord, int lod, int4 color);
-void __ovld write_imageui(write_only image1d_t image, int coord, int lod, uint4 color);
-
-void __ovld write_imagef(write_only image1d_array_t image_array, int2 coord, int lod, float4 color);
-void __ovld write_imagei(write_only image1d_array_t image_array, int2 coord, int lod, int4 color);
-void __ovld write_imageui(write_only image1d_array_t image_array, int2 coord, int lod, uint4 color);
-
-void __ovld write_imagef(write_only image2d_t image, int2 coord, int lod, float4 color);
-void __ovld write_imagei(write_only image2d_t image, int2 coord, int lod, int4 color);
-void __ovld write_imageui(write_only image2d_t image, int2 coord, int lod, uint4 color);
-
-void __ovld write_imagef(write_only image2d_array_t image_array, int4 coord, int lod, float4 color);
-void __ovld write_imagei(write_only image2d_array_t image_array, int4 coord, int lod, int4 color);
-void __ovld write_imageui(write_only image2d_array_t image_array, int4 coord, int lod, uint4 color);
-
-void __ovld write_imagef(write_only image2d_depth_t image, int2 coord, int lod, float color);
-void __ovld write_imagef(write_only image2d_array_depth_t image, int4 coord, int lod, float color);
-
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imagef(write_only image3d_t image, int4 coord, int lod, float4 color);
-void __ovld write_imagei(write_only image3d_t image, int4 coord, int lod, int4 color);
-void __ovld write_imageui(write_only image3d_t image, int4 coord, int lod, uint4 color);
-#endif
-#endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// Image write functions for half4 type
-#ifdef cl_khr_fp16
-void __ovld write_imageh(write_only image1d_t image, int coord, half4 color);
-void __ovld write_imageh(write_only image2d_t image, int2 coord, half4 color);
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imageh(write_only image3d_t image, int4 coord, half4 color);
-#endif
-void __ovld write_imageh(write_only image1d_array_t image, int2 coord, half4 color);
-void __ovld write_imageh(write_only image2d_array_t image, int4 coord, half4 color);
-void __ovld write_imageh(write_only image1d_buffer_t image, int coord, half4 color);
-#endif //cl_khr_fp16
-
-// Image write functions for read_write images
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void __ovld write_imagef(read_write image2d_t image, int2 coord, float4 color);
-void __ovld write_imagei(read_write image2d_t image, int2 coord, int4 color);
-void __ovld write_imageui(read_write image2d_t image, int2 coord, uint4 color);
-
-void __ovld write_imagef(read_write image2d_array_t image_array, int4 coord, float4 color);
-void __ovld write_imagei(read_write image2d_array_t image_array, int4 coord, int4 color);
-void __ovld write_imageui(read_write image2d_array_t image_array, int4 coord, uint4 color);
-
-void __ovld write_imagef(read_write image1d_t image, int coord, float4 color);
-void __ovld write_imagei(read_write image1d_t image, int coord, int4 color);
-void __ovld write_imageui(read_write image1d_t image, int coord, uint4 color);
-
-void __ovld write_imagef(read_write image1d_buffer_t image, int coord, float4 color);
-void __ovld write_imagei(read_write image1d_buffer_t image, int coord, int4 color);
-void __ovld write_imageui(read_write image1d_buffer_t image, int coord, uint4 color);
-
-void __ovld write_imagef(read_write image1d_array_t image_array, int2 coord, float4 color);
-void __ovld write_imagei(read_write image1d_array_t image_array, int2 coord, int4 color);
-void __ovld write_imageui(read_write image1d_array_t image_array, int2 coord, uint4 color);
-
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imagef(read_write image3d_t image, int4 coord, float4 color);
-void __ovld write_imagei(read_write image3d_t image, int4 coord, int4 color);
-void __ovld write_imageui(read_write image3d_t image, int4 coord, uint4 color);
-#endif
-
-#ifdef cl_khr_depth_images
-void __ovld write_imagef(read_write image2d_depth_t image, int2 coord, float color);
-void __ovld write_imagef(read_write image2d_array_depth_t image, int4 coord, float color);
-#endif //cl_khr_depth_images
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifdef cl_khr_mipmap_image
-void __ovld write_imagef(read_write image1d_t image, int coord, int lod, float4 color);
-void __ovld write_imagei(read_write image1d_t image, int coord, int lod, int4 color);
-void __ovld write_imageui(read_write image1d_t image, int coord, int lod, uint4 color);
-
-void __ovld write_imagef(read_write image1d_array_t image_array, int2 coord, int lod, float4 color);
-void __ovld write_imagei(read_write image1d_array_t image_array, int2 coord, int lod, int4 color);
-void __ovld write_imageui(read_write image1d_array_t image_array, int2 coord, int lod, uint4 color);
-
-void __ovld write_imagef(read_write image2d_t image, int2 coord, int lod, float4 color);
-void __ovld write_imagei(read_write image2d_t image, int2 coord, int lod, int4 color);
-void __ovld write_imageui(read_write image2d_t image, int2 coord, int lod, uint4 color);
-
-void __ovld write_imagef(read_write image2d_array_t image_array, int4 coord, int lod, float4 color);
-void __ovld write_imagei(read_write image2d_array_t image_array, int4 coord, int lod, int4 color);
-void __ovld write_imageui(read_write image2d_array_t image_array, int4 coord, int lod, uint4 color);
-
-void __ovld write_imagef(read_write image2d_depth_t image, int2 coord, int lod, float color);
-void __ovld write_imagef(read_write image2d_array_depth_t image, int4 coord, int lod, float color);
-
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imagef(read_write image3d_t image, int4 coord, int lod, float4 color);
-void __ovld write_imagei(read_write image3d_t image, int4 coord, int lod, int4 color);
-void __ovld write_imageui(read_write image3d_t image, int4 coord, int lod, uint4 color);
-#endif
-#endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// Image write functions for half4 type
-#ifdef cl_khr_fp16
-void __ovld write_imageh(read_write image1d_t image, int coord, half4 color);
-void __ovld write_imageh(read_write image2d_t image, int2 coord, half4 color);
-#ifdef cl_khr_3d_image_writes
-void __ovld write_imageh(read_write image3d_t image, int4 coord, half4 color);
-#endif
-void __ovld write_imageh(read_write image1d_array_t image, int2 coord, half4 color);
-void __ovld write_imageh(read_write image2d_array_t image, int4 coord, half4 color);
-void __ovld write_imageh(read_write image1d_buffer_t image, int coord, half4 color);
-#endif //cl_khr_fp16
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// Note: In OpenCL v1.0/1.1/1.2, image argument of image query builtin functions does not have
-// access qualifier, which by default assume read_only access qualifier. Image query builtin
-// functions with write_only image argument should also be declared.
-
-/**
- * Return the image width in pixels.
- *
-  */
-int __ovld __cnfn get_image_width(read_only image1d_t image);
-int __ovld __cnfn get_image_width(read_only image1d_buffer_t image);
-int __ovld __cnfn get_image_width(read_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_width(read_only image3d_t image);
-#endif
-int __ovld __cnfn get_image_width(read_only image1d_array_t image);
-int __ovld __cnfn get_image_width(read_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_width(read_only image2d_depth_t image);
-int __ovld __cnfn get_image_width(read_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_width(read_only image2d_msaa_t image);
-int __ovld __cnfn get_image_width(read_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_width(read_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_width(read_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-int __ovld __cnfn get_image_width(write_only image1d_t image);
-int __ovld __cnfn get_image_width(write_only image1d_buffer_t image);
-int __ovld __cnfn get_image_width(write_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_width(write_only image3d_t image);
-#endif
-int __ovld __cnfn get_image_width(write_only image1d_array_t image);
-int __ovld __cnfn get_image_width(write_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_width(write_only image2d_depth_t image);
-int __ovld __cnfn get_image_width(write_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_width(write_only image2d_msaa_t image);
-int __ovld __cnfn get_image_width(write_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_width(write_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_width(write_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __cnfn get_image_width(read_write image1d_t image);
-int __ovld __cnfn get_image_width(read_write image1d_buffer_t image);
-int __ovld __cnfn get_image_width(read_write image2d_t image);
-int __ovld __cnfn get_image_width(read_write image3d_t image);
-int __ovld __cnfn get_image_width(read_write image1d_array_t image);
-int __ovld __cnfn get_image_width(read_write image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_width(read_write image2d_depth_t image);
-int __ovld __cnfn get_image_width(read_write image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_width(read_write image2d_msaa_t image);
-int __ovld __cnfn get_image_width(read_write image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_width(read_write image2d_array_msaa_t image);
-int __ovld __cnfn get_image_width(read_write image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the image height in pixels.
- */
-int __ovld __cnfn get_image_height(read_only image2d_t image);
-int __ovld __cnfn get_image_height(read_only image3d_t image);
-int __ovld __cnfn get_image_height(read_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_height(read_only image2d_depth_t image);
-int __ovld __cnfn get_image_height(read_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_height(read_only image2d_msaa_t image);
-int __ovld __cnfn get_image_height(read_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_height(read_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_height(read_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-int __ovld __cnfn get_image_height(write_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_height(write_only image3d_t image);
-#endif
-int __ovld __cnfn get_image_height(write_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_height(write_only image2d_depth_t image);
-int __ovld __cnfn get_image_height(write_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_height(write_only image2d_msaa_t image);
-int __ovld __cnfn get_image_height(write_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_height(write_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_height(write_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __cnfn get_image_height(read_write image2d_t image);
-int __ovld __cnfn get_image_height(read_write image3d_t image);
-int __ovld __cnfn get_image_height(read_write image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_height(read_write image2d_depth_t image);
-int __ovld __cnfn get_image_height(read_write image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_height(read_write image2d_msaa_t image);
-int __ovld __cnfn get_image_height(read_write image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_height(read_write image2d_array_msaa_t image);
-int __ovld __cnfn get_image_height(read_write image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the image depth in pixels.
- */
-int __ovld __cnfn get_image_depth(read_only image3d_t image);
-
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_depth(write_only image3d_t image);
-#endif
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __cnfn get_image_depth(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL Extension v2.0 s9.18 - Mipmaps
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#ifdef cl_khr_mipmap_image
-/**
- * Return the image miplevels.
- */
-
-int __ovld get_image_num_mip_levels(read_only image1d_t image);
-int __ovld get_image_num_mip_levels(read_only image2d_t image);
-int __ovld get_image_num_mip_levels(read_only image3d_t image);
-
-int __ovld get_image_num_mip_levels(write_only image1d_t image);
-int __ovld get_image_num_mip_levels(write_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld get_image_num_mip_levels(write_only image3d_t image);
-#endif
-
-int __ovld get_image_num_mip_levels(read_write image1d_t image);
-int __ovld get_image_num_mip_levels(read_write image2d_t image);
-int __ovld get_image_num_mip_levels(read_write image3d_t image);
-
-int __ovld get_image_num_mip_levels(read_only image1d_array_t image);
-int __ovld get_image_num_mip_levels(read_only image2d_array_t image);
-int __ovld get_image_num_mip_levels(read_only image2d_array_depth_t image);
-int __ovld get_image_num_mip_levels(read_only image2d_depth_t image);
-
-int __ovld get_image_num_mip_levels(write_only image1d_array_t image);
-int __ovld get_image_num_mip_levels(write_only image2d_array_t image);
-int __ovld get_image_num_mip_levels(write_only image2d_array_depth_t image);
-int __ovld get_image_num_mip_levels(write_only image2d_depth_t image);
-
-int __ovld get_image_num_mip_levels(read_write image1d_array_t image);
-int __ovld get_image_num_mip_levels(read_write image2d_array_t image);
-int __ovld get_image_num_mip_levels(read_write image2d_array_depth_t image);
-int __ovld get_image_num_mip_levels(read_write image2d_depth_t image);
-
-#endif //cl_khr_mipmap_image
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the channel data type. Valid values are:
- * CLK_SNORM_INT8
- * CLK_SNORM_INT16
- * CLK_UNORM_INT8
- * CLK_UNORM_INT16
- * CLK_UNORM_SHORT_565
- * CLK_UNORM_SHORT_555
- * CLK_UNORM_SHORT_101010
- * CLK_SIGNED_INT8
- * CLK_SIGNED_INT16
- * CLK_SIGNED_INT32
- * CLK_UNSIGNED_INT8
- * CLK_UNSIGNED_INT16
- * CLK_UNSIGNED_INT32
- * CLK_HALF_FLOAT
- * CLK_FLOAT
- */
-
-int __ovld __cnfn get_image_channel_data_type(read_only image1d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image3d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image1d_array_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(read_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-int __ovld __cnfn get_image_channel_data_type(write_only image1d_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_channel_data_type(write_only image3d_t image);
-#endif
-int __ovld __cnfn get_image_channel_data_type(write_only image1d_array_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(write_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __cnfn get_image_channel_data_type(read_write image1d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image3d_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image1d_array_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_data_type(read_write image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the image channel order. Valid values are:
- * CLK_A
- * CLK_R
- * CLK_Rx
- * CLK_RG
- * CLK_RGx
- * CLK_RA
- * CLK_RGB
- * CLK_RGBx
- * CLK_RGBA
- * CLK_ARGB
- * CLK_BGRA
- * CLK_INTENSITY
- * CLK_LUMINANCE
- */
-
-int __ovld __cnfn get_image_channel_order(read_only image1d_t image);
-int __ovld __cnfn get_image_channel_order(read_only image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_t image);
-int __ovld __cnfn get_image_channel_order(read_only image3d_t image);
-int __ovld __cnfn get_image_channel_order(read_only image1d_array_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_order(read_only image2d_depth_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_order(read_only image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_order(read_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-int __ovld __cnfn get_image_channel_order(write_only image1d_t image);
-int __ovld __cnfn get_image_channel_order(write_only image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_t image);
-#ifdef cl_khr_3d_image_writes
-int __ovld __cnfn get_image_channel_order(write_only image3d_t image);
-#endif
-int __ovld __cnfn get_image_channel_order(write_only image1d_array_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_order(write_only image2d_depth_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_order(write_only image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_order(write_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __cnfn get_image_channel_order(read_write image1d_t image);
-int __ovld __cnfn get_image_channel_order(read_write image1d_buffer_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_t image);
-int __ovld __cnfn get_image_channel_order(read_write image3d_t image);
-int __ovld __cnfn get_image_channel_order(read_write image1d_array_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_array_t image);
-#ifdef cl_khr_depth_images
-int __ovld __cnfn get_image_channel_order(read_write image2d_depth_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_array_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld __cnfn get_image_channel_order(read_write image2d_msaa_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_msaa_depth_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_t image);
-int __ovld __cnfn get_image_channel_order(read_write image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the 2D image width and height as an int2
- * type. The width is returned in the x component, and
- * the height in the y component.
- */
-int2 __ovld __cnfn get_image_dim(read_only image2d_t image);
-int2 __ovld __cnfn get_image_dim(read_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int2 __ovld __cnfn get_image_dim(read_only image2d_array_depth_t image);
-int2 __ovld __cnfn get_image_dim(read_only image2d_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int2 __ovld __cnfn get_image_dim(read_only image2d_msaa_t image);
-int2 __ovld __cnfn get_image_dim(read_only image2d_msaa_depth_t image);
-int2 __ovld __cnfn get_image_dim(read_only image2d_array_msaa_t image);
-int2 __ovld __cnfn get_image_dim(read_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-int2 __ovld __cnfn get_image_dim(write_only image2d_t image);
-int2 __ovld __cnfn get_image_dim(write_only image2d_array_t image);
-#ifdef cl_khr_depth_images
-int2 __ovld __cnfn get_image_dim(write_only image2d_array_depth_t image);
-int2 __ovld __cnfn get_image_dim(write_only image2d_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int2 __ovld __cnfn get_image_dim(write_only image2d_msaa_t image);
-int2 __ovld __cnfn get_image_dim(write_only image2d_msaa_depth_t image);
-int2 __ovld __cnfn get_image_dim(write_only image2d_array_msaa_t image);
-int2 __ovld __cnfn get_image_dim(write_only image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int2 __ovld __cnfn get_image_dim(read_write image2d_t image);
-int2 __ovld __cnfn get_image_dim(read_write image2d_array_t image);
-#ifdef cl_khr_depth_images
-int2 __ovld __cnfn get_image_dim(read_write image2d_array_depth_t image);
-int2 __ovld __cnfn get_image_dim(read_write image2d_depth_t image);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-int2 __ovld __cnfn get_image_dim(read_write image2d_msaa_t image);
-int2 __ovld __cnfn get_image_dim(read_write image2d_msaa_depth_t image);
-int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_t image);
-int2 __ovld __cnfn get_image_dim(read_write image2d_array_msaa_depth_t image);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the 3D image width, height, and depth as an
- * int4 type. The width is returned in the x
- * component, height in the y component, depth in the z
- * component and the w component is 0.
- */
-int4 __ovld __cnfn get_image_dim(read_only image3d_t image);
-#ifdef cl_khr_3d_image_writes
-int4 __ovld __cnfn get_image_dim(write_only image3d_t image);
-#endif
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int4 __ovld __cnfn get_image_dim(read_write image3d_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
- * Return the image array size.
- */
-
-size_t __ovld __cnfn get_image_array_size(read_only image1d_array_t image_array);
-size_t __ovld __cnfn get_image_array_size(read_only image2d_array_t image_array);
-#ifdef cl_khr_depth_images
-size_t __ovld __cnfn get_image_array_size(read_only image2d_array_depth_t image_array);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-size_t __ovld __cnfn get_image_array_size(read_only image2d_array_msaa_t image_array);
-size_t __ovld __cnfn get_image_array_size(read_only image2d_array_msaa_depth_t image_array);
-#endif //cl_khr_gl_msaa_sharing
-
-size_t __ovld __cnfn get_image_array_size(write_only image1d_array_t image_array);
-size_t __ovld __cnfn get_image_array_size(write_only image2d_array_t image_array);
-#ifdef cl_khr_depth_images
-size_t __ovld __cnfn get_image_array_size(write_only image2d_array_depth_t image_array);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-size_t __ovld __cnfn get_image_array_size(write_only image2d_array_msaa_t image_array);
-size_t __ovld __cnfn get_image_array_size(write_only image2d_array_msaa_depth_t image_array);
-#endif //cl_khr_gl_msaa_sharing
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-size_t __ovld __cnfn get_image_array_size(read_write image1d_array_t image_array);
-size_t __ovld __cnfn get_image_array_size(read_write image2d_array_t image_array);
-#ifdef cl_khr_depth_images
-size_t __ovld __cnfn get_image_array_size(read_write image2d_array_depth_t image_array);
-#endif //cl_khr_depth_images
-#if defined(cl_khr_gl_msaa_sharing)
-size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_t image_array);
-size_t __ovld __cnfn get_image_array_size(read_write image2d_array_msaa_depth_t image_array);
-#endif //cl_khr_gl_msaa_sharing
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-/**
-* Return the number of samples associated with image
-*/
-#if defined(cl_khr_gl_msaa_sharing)
-int __ovld get_image_num_samples(read_only image2d_msaa_t image);
-int __ovld get_image_num_samples(read_only image2d_msaa_depth_t image);
-int __ovld get_image_num_samples(read_only image2d_array_msaa_t image);
-int __ovld get_image_num_samples(read_only image2d_array_msaa_depth_t image);
-
-int __ovld get_image_num_samples(write_only image2d_msaa_t image);
-int __ovld get_image_num_samples(write_only image2d_msaa_depth_t image);
-int __ovld get_image_num_samples(write_only image2d_array_msaa_t image);
-int __ovld get_image_num_samples(write_only image2d_array_msaa_depth_t image);
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld get_image_num_samples(read_write image2d_msaa_t image);
-int __ovld get_image_num_samples(read_write image2d_msaa_depth_t image);
-int __ovld get_image_num_samples(read_write image2d_array_msaa_t image);
-int __ovld get_image_num_samples(read_write image2d_array_msaa_depth_t image);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-#endif
-
-// OpenCL v2.0 s6.13.15 - Work-group Functions
-
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-int __ovld __conv work_group_all(int predicate);
-int __ovld __conv work_group_any(int predicate);
-
-#ifdef cl_khr_fp16
-half __ovld __conv work_group_broadcast(half a, size_t local_id);
-half __ovld __conv work_group_broadcast(half a, size_t x, size_t y);
-half __ovld __conv work_group_broadcast(half a, size_t x, size_t y, size_t z);
-#endif
-int __ovld __conv work_group_broadcast(int a, size_t local_id);
-int __ovld __conv work_group_broadcast(int a, size_t x, size_t y);
-int __ovld __conv work_group_broadcast(int a, size_t x, size_t y, size_t z);
-uint __ovld __conv work_group_broadcast(uint a, size_t local_id);
-uint __ovld __conv work_group_broadcast(uint a, size_t x, size_t y);
-uint __ovld __conv work_group_broadcast(uint a, size_t x, size_t y, size_t z);
-long __ovld __conv work_group_broadcast(long a, size_t local_id);
-long __ovld __conv work_group_broadcast(long a, size_t x, size_t y);
-long __ovld __conv work_group_broadcast(long a, size_t x, size_t y, size_t z);
-ulong __ovld __conv work_group_broadcast(ulong a, size_t local_id);
-ulong __ovld __conv work_group_broadcast(ulong a, size_t x, size_t y);
-ulong __ovld __conv work_group_broadcast(ulong a, size_t x, size_t y, size_t z);
-float __ovld __conv work_group_broadcast(float a, size_t local_id);
-float __ovld __conv work_group_broadcast(float a, size_t x, size_t y);
-float __ovld __conv work_group_broadcast(float a, size_t x, size_t y, size_t z);
-#ifdef cl_khr_fp64
-double __ovld __conv work_group_broadcast(double a, size_t local_id);
-double __ovld __conv work_group_broadcast(double a, size_t x, size_t y);
-double __ovld __conv work_group_broadcast(double a, size_t x, size_t y, size_t z);
-#endif //cl_khr_fp64
-
-#ifdef cl_khr_fp16
-half __ovld __conv work_group_reduce_add(half x);
-half __ovld __conv work_group_reduce_min(half x);
-half __ovld __conv work_group_reduce_max(half x);
-half __ovld __conv work_group_scan_exclusive_add(half x);
-half __ovld __conv work_group_scan_exclusive_min(half x);
-half __ovld __conv work_group_scan_exclusive_max(half x);
-half __ovld __conv work_group_scan_inclusive_add(half x);
-half __ovld __conv work_group_scan_inclusive_min(half x);
-half __ovld __conv work_group_scan_inclusive_max(half x);
-#endif
-int __ovld __conv work_group_reduce_add(int x);
-int __ovld __conv work_group_reduce_min(int x);
-int __ovld __conv work_group_reduce_max(int x);
-int __ovld __conv work_group_scan_exclusive_add(int x);
-int __ovld __conv work_group_scan_exclusive_min(int x);
-int __ovld __conv work_group_scan_exclusive_max(int x);
-int __ovld __conv work_group_scan_inclusive_add(int x);
-int __ovld __conv work_group_scan_inclusive_min(int x);
-int __ovld __conv work_group_scan_inclusive_max(int x);
-uint __ovld __conv work_group_reduce_add(uint x);
-uint __ovld __conv work_group_reduce_min(uint x);
-uint __ovld __conv work_group_reduce_max(uint x);
-uint __ovld __conv work_group_scan_exclusive_add(uint x);
-uint __ovld __conv work_group_scan_exclusive_min(uint x);
-uint __ovld __conv work_group_scan_exclusive_max(uint x);
-uint __ovld __conv work_group_scan_inclusive_add(uint x);
-uint __ovld __conv work_group_scan_inclusive_min(uint x);
-uint __ovld __conv work_group_scan_inclusive_max(uint x);
-long __ovld __conv work_group_reduce_add(long x);
-long __ovld __conv work_group_reduce_min(long x);
-long __ovld __conv work_group_reduce_max(long x);
-long __ovld __conv work_group_scan_exclusive_add(long x);
-long __ovld __conv work_group_scan_exclusive_min(long x);
-long __ovld __conv work_group_scan_exclusive_max(long x);
-long __ovld __conv work_group_scan_inclusive_add(long x);
-long __ovld __conv work_group_scan_inclusive_min(long x);
-long __ovld __conv work_group_scan_inclusive_max(long x);
-ulong __ovld __conv work_group_reduce_add(ulong x);
-ulong __ovld __conv work_group_reduce_min(ulong x);
-ulong __ovld __conv work_group_reduce_max(ulong x);
-ulong __ovld __conv work_group_scan_exclusive_add(ulong x);
-ulong __ovld __conv work_group_scan_exclusive_min(ulong x);
-ulong __ovld __conv work_group_scan_exclusive_max(ulong x);
-ulong __ovld __conv work_group_scan_inclusive_add(ulong x);
-ulong __ovld __conv work_group_scan_inclusive_min(ulong x);
-ulong __ovld __conv work_group_scan_inclusive_max(ulong x);
-float __ovld __conv work_group_reduce_add(float x);
-float __ovld __conv work_group_reduce_min(float x);
-float __ovld __conv work_group_reduce_max(float x);
-float __ovld __conv work_group_scan_exclusive_add(float x);
-float __ovld __conv work_group_scan_exclusive_min(float x);
-float __ovld __conv work_group_scan_exclusive_max(float x);
-float __ovld __conv work_group_scan_inclusive_add(float x);
-float __ovld __conv work_group_scan_inclusive_min(float x);
-float __ovld __conv work_group_scan_inclusive_max(float x);
-#ifdef cl_khr_fp64
-double __ovld __conv work_group_reduce_add(double x);
-double __ovld __conv work_group_reduce_min(double x);
-double __ovld __conv work_group_reduce_max(double x);
-double __ovld __conv work_group_scan_exclusive_add(double x);
-double __ovld __conv work_group_scan_exclusive_min(double x);
-double __ovld __conv work_group_scan_exclusive_max(double x);
-double __ovld __conv work_group_scan_inclusive_add(double x);
-double __ovld __conv work_group_scan_inclusive_min(double x);
-double __ovld __conv work_group_scan_inclusive_max(double x);
-#endif //cl_khr_fp64
-
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL v2.0 s6.13.16 - Pipe Functions
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-bool __ovld is_valid_reserve_id(reserve_id_t reserve_id);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-
-// OpenCL v2.0 s6.13.17 - Enqueue Kernels
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-ndrange_t __ovld ndrange_1D(size_t);
-ndrange_t __ovld ndrange_1D(size_t, size_t);
-ndrange_t __ovld ndrange_1D(size_t, size_t, size_t);
-
-ndrange_t __ovld ndrange_2D(const size_t[2]);
-ndrange_t __ovld ndrange_2D(const size_t[2], const size_t[2]);
-ndrange_t __ovld ndrange_2D(const size_t[2], const size_t[2], const size_t[2]);
-
-ndrange_t __ovld ndrange_3D(const size_t[3]);
-ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3]);
-ndrange_t __ovld ndrange_3D(const size_t[3], const size_t[3], const size_t[3]);
-
-int __ovld enqueue_marker(queue_t, uint, const __private clk_event_t*, __private clk_event_t*);
-
-void __ovld retain_event(clk_event_t);
-
-void __ovld release_event(clk_event_t);
-
-clk_event_t __ovld create_user_event(void);
-
-void __ovld set_user_event_status(clk_event_t e, int state);
-
-bool __ovld is_valid_event (clk_event_t event);
-
-void __ovld capture_event_profiling_info(clk_event_t, clk_profiling_info, __global void* value);
-
-queue_t __ovld get_default_queue(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-// OpenCL Extension v2.0 s9.17 - Sub-groups
-
-#if defined(cl_intel_subgroups) || defined(cl_khr_subgroups)
-// Shared Sub Group Functions
-uint    __ovld get_sub_group_size(void);
-uint    __ovld get_max_sub_group_size(void);
-uint    __ovld get_num_sub_groups(void);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-uint    __ovld get_enqueued_num_sub_groups(void);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-uint    __ovld get_sub_group_id(void);
-uint    __ovld get_sub_group_local_id(void);
-
-void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags);
-#if __OPENCL_C_VERSION__ >= CL_VERSION_2_0
-void    __ovld __conv sub_group_barrier(cl_mem_fence_flags flags, memory_scope scope);
-#endif //__OPENCL_C_VERSION__ >= CL_VERSION_2_0
-
-int     __ovld __conv sub_group_all(int predicate);
-int     __ovld __conv sub_group_any(int predicate);
-
-int     __ovld __conv sub_group_broadcast(int   x, uint sub_group_local_id);
-uint    __ovld __conv sub_group_broadcast(uint  x, uint sub_group_local_id);
-long    __ovld __conv sub_group_broadcast(long  x, uint sub_group_local_id);
-ulong   __ovld __conv sub_group_broadcast(ulong x, uint sub_group_local_id);
-float   __ovld __conv sub_group_broadcast(float x, uint sub_group_local_id);
-
-int     __ovld __conv sub_group_reduce_add(int   x);
-uint    __ovld __conv sub_group_reduce_add(uint  x);
-long    __ovld __conv sub_group_reduce_add(long  x);
-ulong   __ovld __conv sub_group_reduce_add(ulong x);
-float   __ovld __conv sub_group_reduce_add(float x);
-int     __ovld __conv sub_group_reduce_min(int   x);
-uint    __ovld __conv sub_group_reduce_min(uint  x);
-long    __ovld __conv sub_group_reduce_min(long  x);
-ulong   __ovld __conv sub_group_reduce_min(ulong x);
-float   __ovld __conv sub_group_reduce_min(float x);
-int     __ovld __conv sub_group_reduce_max(int   x);
-uint    __ovld __conv sub_group_reduce_max(uint  x);
-long    __ovld __conv sub_group_reduce_max(long  x);
-ulong   __ovld __conv sub_group_reduce_max(ulong x);
-float   __ovld __conv sub_group_reduce_max(float x);
-
-int     __ovld __conv sub_group_scan_exclusive_add(int   x);
-uint    __ovld __conv sub_group_scan_exclusive_add(uint  x);
-long    __ovld __conv sub_group_scan_exclusive_add(long  x);
-ulong   __ovld __conv sub_group_scan_exclusive_add(ulong x);
-float   __ovld __conv sub_group_scan_exclusive_add(float x);
-int     __ovld __conv sub_group_scan_exclusive_min(int   x);
-uint    __ovld __conv sub_group_scan_exclusive_min(uint  x);
-long    __ovld __conv sub_group_scan_exclusive_min(long  x);
-ulong   __ovld __conv sub_group_scan_exclusive_min(ulong x);
-float   __ovld __conv sub_group_scan_exclusive_min(float x);
-int     __ovld __conv sub_group_scan_exclusive_max(int   x);
-uint    __ovld __conv sub_group_scan_exclusive_max(uint  x);
-long    __ovld __conv sub_group_scan_exclusive_max(long  x);
-ulong   __ovld __conv sub_group_scan_exclusive_max(ulong x);
-float   __ovld __conv sub_group_scan_exclusive_max(float x);
-
-int     __ovld __conv sub_group_scan_inclusive_add(int   x);
-uint    __ovld __conv sub_group_scan_inclusive_add(uint  x);
-long    __ovld __conv sub_group_scan_inclusive_add(long  x);
-ulong   __ovld __conv sub_group_scan_inclusive_add(ulong x);
-float   __ovld __conv sub_group_scan_inclusive_add(float x);
-int     __ovld __conv sub_group_scan_inclusive_min(int   x);
-uint    __ovld __conv sub_group_scan_inclusive_min(uint  x);
-long    __ovld __conv sub_group_scan_inclusive_min(long  x);
-ulong   __ovld __conv sub_group_scan_inclusive_min(ulong x);
-float   __ovld __conv sub_group_scan_inclusive_min(float x);
-int     __ovld __conv sub_group_scan_inclusive_max(int   x);
-uint    __ovld __conv sub_group_scan_inclusive_max(uint  x);
-long    __ovld __conv sub_group_scan_inclusive_max(long  x);
-ulong   __ovld __conv sub_group_scan_inclusive_max(ulong x);
-float   __ovld __conv sub_group_scan_inclusive_max(float x);
-
-#ifdef cl_khr_fp16
-half    __ovld __conv sub_group_broadcast(half x, uint sub_group_local_id);
-half    __ovld __conv sub_group_reduce_add(half x);
-half    __ovld __conv sub_group_reduce_min(half x);
-half    __ovld __conv sub_group_reduce_max(half x);
-half    __ovld __conv sub_group_scan_exclusive_add(half x);
-half    __ovld __conv sub_group_scan_exclusive_min(half x);
-half    __ovld __conv sub_group_scan_exclusive_max(half x);
-half    __ovld __conv sub_group_scan_inclusive_add(half x);
-half    __ovld __conv sub_group_scan_inclusive_min(half x);
-half    __ovld __conv sub_group_scan_inclusive_max(half x);
-#endif //cl_khr_fp16
-
-#ifdef cl_khr_fp64
-double  __ovld __conv sub_group_broadcast(double x, uint sub_group_local_id);
-double  __ovld __conv sub_group_reduce_add(double x);
-double  __ovld __conv sub_group_reduce_min(double x);
-double  __ovld __conv sub_group_reduce_max(double x);
-double  __ovld __conv sub_group_scan_exclusive_add(double x);
-double  __ovld __conv sub_group_scan_exclusive_min(double x);
-double  __ovld __conv sub_group_scan_exclusive_max(double x);
-double  __ovld __conv sub_group_scan_inclusive_add(double x);
-double  __ovld __conv sub_group_scan_inclusive_min(double x);
-double  __ovld __conv sub_group_scan_inclusive_max(double x);
-#endif //cl_khr_fp64
-
-#endif //cl_khr_subgroups cl_intel_subgroups
-
-#if defined(cl_intel_subgroups)
-// Intel-Specific Sub Group Functions
-float   __ovld __conv intel_sub_group_shuffle( float  x, uint c );
-float2  __ovld __conv intel_sub_group_shuffle( float2 x, uint c );
-float3  __ovld __conv intel_sub_group_shuffle( float3 x, uint c );
-float4  __ovld __conv intel_sub_group_shuffle( float4 x, uint c );
-float8  __ovld __conv intel_sub_group_shuffle( float8 x, uint c );
-float16 __ovld __conv intel_sub_group_shuffle( float16 x, uint c );
-
-int     __ovld __conv intel_sub_group_shuffle( int  x, uint c );
-int2    __ovld __conv intel_sub_group_shuffle( int2 x, uint c );
-int3    __ovld __conv intel_sub_group_shuffle( int3 x, uint c );
-int4    __ovld __conv intel_sub_group_shuffle( int4 x, uint c );
-int8    __ovld __conv intel_sub_group_shuffle( int8 x, uint c );
-int16   __ovld __conv intel_sub_group_shuffle( int16 x, uint c );
-
-uint    __ovld __conv intel_sub_group_shuffle( uint  x, uint c );
-uint2   __ovld __conv intel_sub_group_shuffle( uint2 x, uint c );
-uint3   __ovld __conv intel_sub_group_shuffle( uint3 x, uint c );
-uint4   __ovld __conv intel_sub_group_shuffle( uint4 x, uint c );
-uint8   __ovld __conv intel_sub_group_shuffle( uint8 x, uint c );
-uint16  __ovld __conv intel_sub_group_shuffle( uint16 x, uint c );
-
-long    __ovld __conv intel_sub_group_shuffle( long x, uint c );
-ulong   __ovld __conv intel_sub_group_shuffle( ulong x, uint c );
-
-float   __ovld __conv intel_sub_group_shuffle_down( float  cur, float  next, uint c );
-float2  __ovld __conv intel_sub_group_shuffle_down( float2 cur, float2 next, uint c );
-float3  __ovld __conv intel_sub_group_shuffle_down( float3 cur, float3 next, uint c );
-float4  __ovld __conv intel_sub_group_shuffle_down( float4 cur, float4 next, uint c );
-float8  __ovld __conv intel_sub_group_shuffle_down( float8 cur, float8 next, uint c );
-float16 __ovld __conv intel_sub_group_shuffle_down( float16 cur, float16 next, uint c );
-
-int     __ovld __conv intel_sub_group_shuffle_down( int  cur, int  next, uint c );
-int2    __ovld __conv intel_sub_group_shuffle_down( int2 cur, int2 next, uint c );
-int3    __ovld __conv intel_sub_group_shuffle_down( int3 cur, int3 next, uint c );
-int4    __ovld __conv intel_sub_group_shuffle_down( int4 cur, int4 next, uint c );
-int8    __ovld __conv intel_sub_group_shuffle_down( int8 cur, int8 next, uint c );
-int16   __ovld __conv intel_sub_group_shuffle_down( int16 cur, int16 next, uint c );
-
-uint    __ovld __conv intel_sub_group_shuffle_down( uint  cur, uint  next, uint c );
-uint2   __ovld __conv intel_sub_group_shuffle_down( uint2 cur, uint2 next, uint c );
-uint3   __ovld __conv intel_sub_group_shuffle_down( uint3 cur, uint3 next, uint c );
-uint4   __ovld __conv intel_sub_group_shuffle_down( uint4 cur, uint4 next, uint c );
-uint8   __ovld __conv intel_sub_group_shuffle_down( uint8 cur, uint8 next, uint c );
-uint16  __ovld __conv intel_sub_group_shuffle_down( uint16 cur, uint16 next, uint c );
-
-long    __ovld __conv intel_sub_group_shuffle_down( long prev, long cur, uint c );
-ulong   __ovld __conv intel_sub_group_shuffle_down( ulong prev, ulong cur, uint c );
-
-float   __ovld __conv intel_sub_group_shuffle_up( float  prev, float  cur, uint c );
-float2  __ovld __conv intel_sub_group_shuffle_up( float2 prev, float2 cur, uint c );
-float3  __ovld __conv intel_sub_group_shuffle_up( float3 prev, float3 cur, uint c );
-float4  __ovld __conv intel_sub_group_shuffle_up( float4 prev, float4 cur, uint c );
-float8  __ovld __conv intel_sub_group_shuffle_up( float8 prev, float8 cur, uint c );
-float16 __ovld __conv intel_sub_group_shuffle_up( float16 prev, float16 cur, uint c );
-
-int     __ovld __conv intel_sub_group_shuffle_up( int  prev, int  cur, uint c );
-int2    __ovld __conv intel_sub_group_shuffle_up( int2 prev, int2 cur, uint c );
-int3    __ovld __conv intel_sub_group_shuffle_up( int3 prev, int3 cur, uint c );
-int4    __ovld __conv intel_sub_group_shuffle_up( int4 prev, int4 cur, uint c );
-int8    __ovld __conv intel_sub_group_shuffle_up( int8 prev, int8 cur, uint c );
-int16   __ovld __conv intel_sub_group_shuffle_up( int16 prev, int16 cur, uint c );
-
-uint    __ovld __conv intel_sub_group_shuffle_up( uint  prev, uint  cur, uint c );
-uint2   __ovld __conv intel_sub_group_shuffle_up( uint2 prev, uint2 cur, uint c );
-uint3   __ovld __conv intel_sub_group_shuffle_up( uint3 prev, uint3 cur, uint c );
-uint4   __ovld __conv intel_sub_group_shuffle_up( uint4 prev, uint4 cur, uint c );
-uint8   __ovld __conv intel_sub_group_shuffle_up( uint8 prev, uint8 cur, uint c );
-uint16  __ovld __conv intel_sub_group_shuffle_up( uint16 prev, uint16 cur, uint c );
-
-long    __ovld __conv intel_sub_group_shuffle_up( long prev, long cur, uint c );
-ulong   __ovld __conv intel_sub_group_shuffle_up( ulong prev, ulong cur, uint c );
-
-float   __ovld __conv intel_sub_group_shuffle_xor( float  x, uint c );
-float2  __ovld __conv intel_sub_group_shuffle_xor( float2 x, uint c );
-float3  __ovld __conv intel_sub_group_shuffle_xor( float3 x, uint c );
-float4  __ovld __conv intel_sub_group_shuffle_xor( float4 x, uint c );
-float8  __ovld __conv intel_sub_group_shuffle_xor( float8 x, uint c );
-float16 __ovld __conv intel_sub_group_shuffle_xor( float16 x, uint c );
-
-int     __ovld __conv intel_sub_group_shuffle_xor( int  x, uint c );
-int2    __ovld __conv intel_sub_group_shuffle_xor( int2 x, uint c );
-int3    __ovld __conv intel_sub_group_shuffle_xor( int3 x, uint c );
-int4    __ovld __conv intel_sub_group_shuffle_xor( int4 x, uint c );
-int8    __ovld __conv intel_sub_group_shuffle_xor( int8 x, uint c );
-int16   __ovld __conv intel_sub_group_shuffle_xor( int16 x, uint c );
-
-uint    __ovld __conv intel_sub_group_shuffle_xor( uint  x, uint c );
-uint2   __ovld __conv intel_sub_group_shuffle_xor( uint2 x, uint c );
-uint3   __ovld __conv intel_sub_group_shuffle_xor( uint3 x, uint c );
-uint4   __ovld __conv intel_sub_group_shuffle_xor( uint4 x, uint c );
-uint8   __ovld __conv intel_sub_group_shuffle_xor( uint8 x, uint c );
-uint16  __ovld __conv intel_sub_group_shuffle_xor( uint16 x, uint c );
-
-long    __ovld __conv intel_sub_group_shuffle_xor( long x, uint c );
-ulong   __ovld __conv intel_sub_group_shuffle_xor( ulong x, uint c );
-
-uint    __ovld __conv intel_sub_group_block_read( read_only image2d_t image, int2 coord );
-uint2   __ovld __conv intel_sub_group_block_read2( read_only image2d_t image, int2 coord );
-uint4   __ovld __conv intel_sub_group_block_read4( read_only image2d_t image, int2 coord );
-uint8   __ovld __conv intel_sub_group_block_read8( read_only image2d_t image, int2 coord );
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-uint    __ovld __conv intel_sub_group_block_read(read_write image2d_t image, int2 coord);
-uint2   __ovld __conv intel_sub_group_block_read2(read_write image2d_t image, int2 coord);
-uint4   __ovld __conv intel_sub_group_block_read4(read_write image2d_t image, int2 coord);
-uint8   __ovld __conv intel_sub_group_block_read8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-uint    __ovld __conv intel_sub_group_block_read( const __global uint* p );
-uint2   __ovld __conv intel_sub_group_block_read2( const __global uint* p );
-uint4   __ovld __conv intel_sub_group_block_read4( const __global uint* p );
-uint8   __ovld __conv intel_sub_group_block_read8( const __global uint* p );
-
-void    __ovld __conv intel_sub_group_block_write(write_only image2d_t image, int2 coord, uint data);
-void    __ovld __conv intel_sub_group_block_write2(write_only image2d_t image, int2 coord, uint2 data);
-void    __ovld __conv intel_sub_group_block_write4(write_only image2d_t image, int2 coord, uint4 data);
-void    __ovld __conv intel_sub_group_block_write8(write_only image2d_t image, int2 coord, uint8 data);
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-void    __ovld __conv intel_sub_group_block_write(read_write image2d_t image, int2 coord, uint data);
-void    __ovld __conv intel_sub_group_block_write2(read_write image2d_t image, int2 coord, uint2 data);
-void    __ovld __conv intel_sub_group_block_write4(read_write image2d_t image, int2 coord, uint4 data);
-void    __ovld __conv intel_sub_group_block_write8(read_write image2d_t image, int2 coord, uint8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-void    __ovld __conv intel_sub_group_block_write( __global uint* p, uint data );
-void    __ovld __conv intel_sub_group_block_write2( __global uint* p, uint2 data );
-void    __ovld __conv intel_sub_group_block_write4( __global uint* p, uint4 data );
-void    __ovld __conv intel_sub_group_block_write8( __global uint* p, uint8 data );
-
-#ifdef cl_khr_fp16
-half    __ovld __conv intel_sub_group_shuffle( half x, uint c );
-half    __ovld __conv intel_sub_group_shuffle_down( half prev, half cur, uint c );
-half    __ovld __conv intel_sub_group_shuffle_up( half prev, half cur, uint c );
-half    __ovld __conv intel_sub_group_shuffle_xor( half x, uint c );
-#endif
-
-#if defined(cl_khr_fp64)
-double  __ovld __conv intel_sub_group_shuffle( double x, uint c );
-double  __ovld __conv intel_sub_group_shuffle_down( double prev, double cur, uint c );
-double  __ovld __conv intel_sub_group_shuffle_up( double prev, double cur, uint c );
-double  __ovld __conv intel_sub_group_shuffle_xor( double x, uint c );
-#endif
-
-#endif //cl_intel_subgroups
-
-#if defined(cl_intel_subgroups_short)
-short       __ovld __conv intel_sub_group_broadcast( short  x, uint sub_group_local_id );
-short2      __ovld __conv intel_sub_group_broadcast( short2 x, uint sub_group_local_id );
-short3      __ovld __conv intel_sub_group_broadcast( short3 x, uint sub_group_local_id );
-short4      __ovld __conv intel_sub_group_broadcast( short4 x, uint sub_group_local_id );
-short8      __ovld __conv intel_sub_group_broadcast( short8 x, uint sub_group_local_id );
-
-ushort      __ovld __conv intel_sub_group_broadcast( ushort  x, uint sub_group_local_id );
-ushort2     __ovld __conv intel_sub_group_broadcast( ushort2 x, uint sub_group_local_id );
-ushort3     __ovld __conv intel_sub_group_broadcast( ushort3 x, uint sub_group_local_id );
-ushort4     __ovld __conv intel_sub_group_broadcast( ushort4 x, uint sub_group_local_id );
-ushort8     __ovld __conv intel_sub_group_broadcast( ushort8 x, uint sub_group_local_id );
-
-short       __ovld __conv intel_sub_group_shuffle( short   x, uint c );
-short2      __ovld __conv intel_sub_group_shuffle( short2  x, uint c );
-short3      __ovld __conv intel_sub_group_shuffle( short3  x, uint c );
-short4      __ovld __conv intel_sub_group_shuffle( short4  x, uint c );
-short8      __ovld __conv intel_sub_group_shuffle( short8  x, uint c );
-short16     __ovld __conv intel_sub_group_shuffle( short16 x, uint c);
-
-ushort      __ovld __conv intel_sub_group_shuffle( ushort   x, uint c );
-ushort2     __ovld __conv intel_sub_group_shuffle( ushort2  x, uint c );
-ushort3     __ovld __conv intel_sub_group_shuffle( ushort3  x, uint c );
-ushort4     __ovld __conv intel_sub_group_shuffle( ushort4  x, uint c );
-ushort8     __ovld __conv intel_sub_group_shuffle( ushort8  x, uint c );
-ushort16    __ovld __conv intel_sub_group_shuffle( ushort16 x, uint c );
-
-short       __ovld __conv intel_sub_group_shuffle_down( short   cur, short   next, uint c );
-short2      __ovld __conv intel_sub_group_shuffle_down( short2  cur, short2  next, uint c );
-short3      __ovld __conv intel_sub_group_shuffle_down( short3  cur, short3  next, uint c );
-short4      __ovld __conv intel_sub_group_shuffle_down( short4  cur, short4  next, uint c );
-short8      __ovld __conv intel_sub_group_shuffle_down( short8  cur, short8  next, uint c );
-short16     __ovld __conv intel_sub_group_shuffle_down( short16 cur, short16 next, uint c );
-
-ushort      __ovld __conv intel_sub_group_shuffle_down( ushort   cur, ushort   next, uint c );
-ushort2     __ovld __conv intel_sub_group_shuffle_down( ushort2  cur, ushort2  next, uint c );
-ushort3     __ovld __conv intel_sub_group_shuffle_down( ushort3  cur, ushort3  next, uint c );
-ushort4     __ovld __conv intel_sub_group_shuffle_down( ushort4  cur, ushort4  next, uint c );
-ushort8     __ovld __conv intel_sub_group_shuffle_down( ushort8  cur, ushort8  next, uint c );
-ushort16    __ovld __conv intel_sub_group_shuffle_down( ushort16 cur, ushort16 next, uint c );
-
-short       __ovld __conv intel_sub_group_shuffle_up( short   cur, short   next, uint c );
-short2      __ovld __conv intel_sub_group_shuffle_up( short2  cur, short2  next, uint c );
-short3      __ovld __conv intel_sub_group_shuffle_up( short3  cur, short3  next, uint c );
-short4      __ovld __conv intel_sub_group_shuffle_up( short4  cur, short4  next, uint c );
-short8      __ovld __conv intel_sub_group_shuffle_up( short8  cur, short8  next, uint c );
-short16     __ovld __conv intel_sub_group_shuffle_up( short16 cur, short16 next, uint c );
-
-ushort      __ovld __conv intel_sub_group_shuffle_up( ushort   cur, ushort   next, uint c );
-ushort2     __ovld __conv intel_sub_group_shuffle_up( ushort2  cur, ushort2  next, uint c );
-ushort3     __ovld __conv intel_sub_group_shuffle_up( ushort3  cur, ushort3  next, uint c );
-ushort4     __ovld __conv intel_sub_group_shuffle_up( ushort4  cur, ushort4  next, uint c );
-ushort8     __ovld __conv intel_sub_group_shuffle_up( ushort8  cur, ushort8  next, uint c );
-ushort16    __ovld __conv intel_sub_group_shuffle_up( ushort16 cur, ushort16 next, uint c );
-
-short       __ovld __conv intel_sub_group_shuffle_xor( short   x, uint c );
-short2      __ovld __conv intel_sub_group_shuffle_xor( short2  x, uint c );
-short3      __ovld __conv intel_sub_group_shuffle_xor( short3  x, uint c );
-short4      __ovld __conv intel_sub_group_shuffle_xor( short4  x, uint c );
-short8      __ovld __conv intel_sub_group_shuffle_xor( short8  x, uint c );
-short16     __ovld __conv intel_sub_group_shuffle_xor( short16 x, uint c );
-
-ushort      __ovld __conv intel_sub_group_shuffle_xor( ushort   x, uint c );
-ushort2     __ovld __conv intel_sub_group_shuffle_xor( ushort2  x, uint c );
-ushort3     __ovld __conv intel_sub_group_shuffle_xor( ushort3  x, uint c );
-ushort4     __ovld __conv intel_sub_group_shuffle_xor( ushort4  x, uint c );
-ushort8     __ovld __conv intel_sub_group_shuffle_xor( ushort8  x, uint c );
-ushort16    __ovld __conv intel_sub_group_shuffle_xor( ushort16 x, uint c );
-
-short       __ovld __conv intel_sub_group_reduce_add( short   x );
-ushort      __ovld __conv intel_sub_group_reduce_add( ushort  x );
-short       __ovld __conv intel_sub_group_reduce_min( short   x );
-ushort      __ovld __conv intel_sub_group_reduce_min( ushort  x );
-short       __ovld __conv intel_sub_group_reduce_max( short   x );
-ushort      __ovld __conv intel_sub_group_reduce_max( ushort  x );
-
-short       __ovld __conv intel_sub_group_scan_exclusive_add( short   x );
-ushort      __ovld __conv intel_sub_group_scan_exclusive_add( ushort  x );
-short       __ovld __conv intel_sub_group_scan_exclusive_min( short   x );
-ushort      __ovld __conv intel_sub_group_scan_exclusive_min( ushort  x );
-short       __ovld __conv intel_sub_group_scan_exclusive_max( short   x );
-ushort      __ovld __conv intel_sub_group_scan_exclusive_max( ushort  x );
-
-short       __ovld __conv intel_sub_group_scan_inclusive_add( short   x );
-ushort      __ovld __conv intel_sub_group_scan_inclusive_add( ushort  x );
-short       __ovld __conv intel_sub_group_scan_inclusive_min( short   x );
-ushort      __ovld __conv intel_sub_group_scan_inclusive_min( ushort  x );
-short       __ovld __conv intel_sub_group_scan_inclusive_max( short   x );
-ushort      __ovld __conv intel_sub_group_scan_inclusive_max( ushort  x );
-
-uint       __ovld __conv intel_sub_group_block_read_ui( read_only image2d_t image, int2 byte_coord );
-uint2      __ovld __conv intel_sub_group_block_read_ui2( read_only image2d_t image, int2 byte_coord );
-uint4      __ovld __conv intel_sub_group_block_read_ui4( read_only image2d_t image, int2 byte_coord );
-uint8      __ovld __conv intel_sub_group_block_read_ui8( read_only image2d_t image, int2 byte_coord );
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-uint       __ovld __conv intel_sub_group_block_read_ui( read_write image2d_t image, int2 byte_coord );
-uint2      __ovld __conv intel_sub_group_block_read_ui2( read_write image2d_t image, int2 byte_coord );
-uint4      __ovld __conv intel_sub_group_block_read_ui4( read_write image2d_t image, int2 byte_coord );
-uint8      __ovld __conv intel_sub_group_block_read_ui8( read_write image2d_t image, int2 byte_coord );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-uint       __ovld __conv intel_sub_group_block_read_ui( const __global uint* p );
-uint2      __ovld __conv intel_sub_group_block_read_ui2( const __global uint* p );
-uint4      __ovld __conv intel_sub_group_block_read_ui4( const __global uint* p );
-uint8      __ovld __conv intel_sub_group_block_read_ui8( const __global uint* p );
-
-void       __ovld __conv intel_sub_group_block_write_ui( read_only image2d_t image, int2 byte_coord, uint data );
-void       __ovld __conv intel_sub_group_block_write_ui2( read_only image2d_t image, int2 byte_coord, uint2 data );
-void       __ovld __conv intel_sub_group_block_write_ui4( read_only image2d_t image, int2 byte_coord, uint4 data );
-void       __ovld __conv intel_sub_group_block_write_ui8( read_only image2d_t image, int2 byte_coord, uint8 data );
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-void       __ovld __conv intel_sub_group_block_write_ui( read_write image2d_t image, int2 byte_coord, uint data );
-void       __ovld __conv intel_sub_group_block_write_ui2( read_write image2d_t image, int2 byte_coord, uint2 data );
-void       __ovld __conv intel_sub_group_block_write_ui4( read_write image2d_t image, int2 byte_coord, uint4 data );
-void       __ovld __conv intel_sub_group_block_write_ui8( read_write image2d_t image, int2 byte_coord, uint8 data );
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-void       __ovld __conv intel_sub_group_block_write_ui( __global uint* p, uint data );
-void       __ovld __conv intel_sub_group_block_write_ui2( __global uint* p, uint2 data );
-void       __ovld __conv intel_sub_group_block_write_ui4( __global uint* p, uint4 data );
-void       __ovld __conv intel_sub_group_block_write_ui8( __global uint* p, uint8 data );
-
-ushort      __ovld __conv intel_sub_group_block_read_us( read_only image2d_t image, int2 coord );
-ushort2     __ovld __conv intel_sub_group_block_read_us2( read_only image2d_t image, int2 coord );
-ushort4     __ovld __conv intel_sub_group_block_read_us4( read_only image2d_t image, int2 coord );
-ushort8     __ovld __conv intel_sub_group_block_read_us8( read_only image2d_t image, int2 coord );
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-ushort      __ovld __conv intel_sub_group_block_read_us(read_write image2d_t image, int2 coord);
-ushort2     __ovld __conv intel_sub_group_block_read_us2(read_write image2d_t image, int2 coord);
-ushort4     __ovld __conv intel_sub_group_block_read_us4(read_write image2d_t image, int2 coord);
-ushort8     __ovld __conv intel_sub_group_block_read_us8(read_write image2d_t image, int2 coord);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-ushort      __ovld __conv intel_sub_group_block_read_us(  const __global ushort* p );
-ushort2     __ovld __conv intel_sub_group_block_read_us2( const __global ushort* p );
-ushort4     __ovld __conv intel_sub_group_block_read_us4( const __global ushort* p );
-ushort8     __ovld __conv intel_sub_group_block_read_us8( const __global ushort* p );
-
-void        __ovld __conv intel_sub_group_block_write_us(write_only image2d_t image, int2 coord, ushort  data);
-void        __ovld __conv intel_sub_group_block_write_us2(write_only image2d_t image, int2 coord, ushort2 data);
-void        __ovld __conv intel_sub_group_block_write_us4(write_only image2d_t image, int2 coord, ushort4 data);
-void        __ovld __conv intel_sub_group_block_write_us8(write_only image2d_t image, int2 coord, ushort8 data);
-
-#if (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-void        __ovld __conv intel_sub_group_block_write_us(read_write image2d_t image, int2 coord, ushort  data);
-void        __ovld __conv intel_sub_group_block_write_us2(read_write image2d_t image, int2 coord, ushort2 data);
-void        __ovld __conv intel_sub_group_block_write_us4(read_write image2d_t image, int2 coord, ushort4 data);
-void        __ovld __conv intel_sub_group_block_write_us8(read_write image2d_t image, int2 coord, ushort8 data);
-#endif // (__OPENCL_C_VERSION__ >= CL_VERSION_2_0)
-
-void        __ovld __conv intel_sub_group_block_write_us(  __global ushort* p, ushort  data );
-void        __ovld __conv intel_sub_group_block_write_us2( __global ushort* p, ushort2 data );
-void        __ovld __conv intel_sub_group_block_write_us4( __global ushort* p, ushort4 data );
-void        __ovld __conv intel_sub_group_block_write_us8( __global ushort* p, ushort8 data );
-#endif // cl_intel_subgroups_short
-
-#ifdef cl_intel_device_side_avc_motion_estimation
-#pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : begin
-
-// MCE built-in functions
-uchar __ovld
-intel_sub_group_avc_mce_get_default_inter_base_multi_reference_penalty(
-    uchar slice_type, uchar qp);
-ulong __ovld intel_sub_group_avc_mce_get_default_inter_shape_penalty(
-    uchar slice_type, uchar qp);
-uchar __ovld intel_sub_group_avc_mce_get_default_inter_direction_penalty(
-    uchar slice_type, uchar qp);
-uint __ovld intel_sub_group_avc_mce_get_default_intra_luma_shape_penalty(
-    uchar slice_type, uchar qp);
-uint2 __ovld
-intel_sub_group_avc_mce_get_default_inter_motion_vector_cost_table(
-    uchar slice_type, uchar qp);
-uchar __ovld intel_sub_group_avc_mce_get_default_intra_luma_mode_penalty(
-    uchar slice_type, uchar qp);
-
-uint2 __ovld intel_sub_group_avc_mce_get_default_high_penalty_cost_table();
-uint2 __ovld intel_sub_group_avc_mce_get_default_medium_penalty_cost_table();
-uint2 __ovld intel_sub_group_avc_mce_get_default_low_penalty_cost_table();
-uint __ovld intel_sub_group_avc_mce_get_default_non_dc_luma_intra_penalty();
-uchar __ovld
-intel_sub_group_avc_mce_get_default_intra_chroma_mode_base_penalty();
-
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_inter_base_multi_reference_penalty(
-    uchar reference_base_penalty, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_inter_shape_penalty(
-    ulong packed_shape_penalty, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_inter_direction_penalty(
-    uchar direction_cost, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_motion_vector_cost_function(
-    ulong packed_cost_center_delta, uint2 packed_cost_table,
-    uchar cost_precision, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_ac_only_haar(
-    intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_source_interlaced_field_polarity(
-    uchar src_field_polarity, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_single_reference_interlaced_field_polarity(
-    uchar ref_field_polarity, intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_mce_set_dual_reference_interlaced_field_polarities(
-    uchar fwd_ref_field_polarity, uchar bwd_ref_field_polarity,
-    intel_sub_group_avc_mce_payload_t payload);
-
-ulong __ovld intel_sub_group_avc_mce_get_motion_vectors(
-    intel_sub_group_avc_mce_result_t result);
-ushort __ovld intel_sub_group_avc_mce_get_inter_distortions(
-    intel_sub_group_avc_mce_result_t result);
-ushort __ovld intel_sub_group_avc_mce_get_best_inter_distortion(
-    intel_sub_group_avc_mce_result_t result);
-uchar __ovld intel_sub_group_avc_mce_get_inter_major_shape(
-    intel_sub_group_avc_mce_result_t result);
-uchar __ovld intel_sub_group_avc_mce_get_inter_minor_shapes(
-    intel_sub_group_avc_mce_result_t result);
-uchar __ovld intel_sub_group_avc_mce_get_inter_directions(
-    intel_sub_group_avc_mce_result_t result);
-uchar __ovld intel_sub_group_avc_mce_get_inter_motion_vector_count(
-    intel_sub_group_avc_mce_result_t result);
-uint __ovld intel_sub_group_avc_mce_get_inter_reference_ids(
-    intel_sub_group_avc_mce_result_t result);
-uchar __ovld
-intel_sub_group_avc_mce_get_inter_reference_interlaced_field_polarities(
-    uint packed_reference_ids, uint packed_reference_parameter_field_polarities,
-    intel_sub_group_avc_mce_result_t result);
-
-// IME built-in functions
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_initialize(
-    ushort2 src_coord, uchar partition_mask, uchar sad_adjustment);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_single_reference(
-    short2 ref_offset, uchar search_window_config,
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_dual_reference(
-    short2 fwd_ref_offset, short2 bwd_ref_offset, uchar search_window_config,
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_max_motion_vector_count(
-    uchar max_motion_vector_count, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_unidirectional_mix_disable(
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_early_search_termination_threshold(
-    uchar threshold, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_weighted_sad(
-    uint packed_sad_weights, intel_sub_group_avc_ime_payload_t payload);
-
-__attribute__((deprecated("If you use the latest Intel driver, please use "
-                          "intel_sub_group_avc_ime_ref_window_size instead",
-                          "intel_sub_group_avc_ime_ref_window_size")))
-ushort2 __ovld
-intel_sub_group_ime_ref_window_size(uchar search_window_config, char dual_ref);
-ushort2 __ovld intel_sub_group_avc_ime_ref_window_size(
-    uchar search_window_config, char dual_ref);
-short2 __ovld intel_sub_group_avc_ime_adjust_ref_offset(
-    short2 ref_offset, ushort2 src_coord, ushort2 ref_window_size,
-    ushort2 image_size);
-
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_evaluate_with_single_reference(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_evaluate_with_dual_reference(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_result_single_reference_streamout_t __ovld
-intel_sub_group_avc_ime_evaluate_with_single_reference_streamout(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_result_dual_reference_streamout_t __ovld
-intel_sub_group_avc_ime_evaluate_with_dual_reference_streamout(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_evaluate_with_single_reference_streamin(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ime_payload_t payload,
-    intel_sub_group_avc_ime_single_reference_streamin_t streamin_components);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_evaluate_with_dual_reference_streamin(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ime_payload_t payload,
-    intel_sub_group_avc_ime_dual_reference_streamin_t streamin_components);
-intel_sub_group_avc_ime_result_single_reference_streamout_t __ovld
-intel_sub_group_avc_ime_evaluate_with_single_reference_streaminout(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ime_payload_t payload,
-    intel_sub_group_avc_ime_single_reference_streamin_t streamin_components);
-intel_sub_group_avc_ime_result_dual_reference_streamout_t __ovld
-intel_sub_group_avc_ime_evaluate_with_dual_reference_streaminout(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ime_payload_t payload,
-    intel_sub_group_avc_ime_dual_reference_streamin_t streamin_components);
-
-intel_sub_group_avc_ime_single_reference_streamin_t __ovld
-intel_sub_group_avc_ime_get_single_reference_streamin(
-    intel_sub_group_avc_ime_result_single_reference_streamout_t result);
-intel_sub_group_avc_ime_dual_reference_streamin_t __ovld
-intel_sub_group_avc_ime_get_dual_reference_streamin(
-    intel_sub_group_avc_ime_result_dual_reference_streamout_t result);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_strip_single_reference_streamout(
-    intel_sub_group_avc_ime_result_single_reference_streamout_t result);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_ime_strip_dual_reference_streamout(
-    intel_sub_group_avc_ime_result_dual_reference_streamout_t result);
-
-uint __ovld intel_sub_group_avc_ime_get_streamout_major_shape_motion_vectors(
-    intel_sub_group_avc_ime_result_single_reference_streamout_t result,
-    uchar major_shape);
-ushort __ovld intel_sub_group_avc_ime_get_streamout_major_shape_distortions(
-    intel_sub_group_avc_ime_result_single_reference_streamout_t result,
-    uchar major_shape);
-uchar __ovld intel_sub_group_avc_ime_get_streamout_major_shape_reference_ids(
-    intel_sub_group_avc_ime_result_single_reference_streamout_t result,
-    uchar major_shape);
-uint __ovld intel_sub_group_avc_ime_get_streamout_major_shape_motion_vectors(
-    intel_sub_group_avc_ime_result_dual_reference_streamout_t result,
-    uchar major_shape, uchar direction);
-ushort __ovld intel_sub_group_avc_ime_get_streamout_major_shape_distortions(
-    intel_sub_group_avc_ime_result_dual_reference_streamout_t result,
-    uchar major_shape, uchar direction);
-uchar __ovld intel_sub_group_avc_ime_get_streamout_major_shape_reference_ids(
-    intel_sub_group_avc_ime_result_dual_reference_streamout_t result,
-    uchar major_shape, uchar direction);
-
-uchar __ovld intel_sub_group_avc_ime_get_border_reached(
-    uchar image_select, intel_sub_group_avc_ime_result_t result);
-uchar __ovld intel_sub_group_avc_ime_get_truncated_search_indication(
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld
-intel_sub_group_avc_ime_get_unidirectional_early_search_termination(
-    intel_sub_group_avc_ime_result_t result);
-uint __ovld intel_sub_group_avc_ime_get_weighting_pattern_minimum_motion_vector(
-    intel_sub_group_avc_ime_result_t result);
-ushort __ovld intel_sub_group_avc_ime_get_weighting_pattern_minimum_distortion(
-    intel_sub_group_avc_ime_result_t result);
-
-// REF built-in functions
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_fme_initialize(
-    ushort2 src_coord, ulong motion_vectors, uchar major_shapes,
-    uchar minor_shapes, uchar directions, uchar pixel_resolution,
-    uchar sad_adjustment);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_bme_initialize(
-    ushort2 src_coord, ulong motion_vectors, uchar major_shapes,
-    uchar minor_shapes, uchar directions, uchar pixel_resolution,
-    uchar bidirectional_weight, uchar sad_adjustment);
-
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_bidirectional_mix_disable(
-    intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_bilinear_filter_enable(
-    intel_sub_group_avc_ref_payload_t payload);
-
-intel_sub_group_avc_ref_result_t __ovld
-intel_sub_group_avc_ref_evaluate_with_single_reference(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_ref_result_t __ovld
-intel_sub_group_avc_ref_evaluate_with_dual_reference(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_ref_result_t __ovld
-intel_sub_group_avc_ref_evaluate_with_multi_reference(
-    read_only image2d_t src_image, uint packed_reference_ids,
-    sampler_t vme_media_sampler, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_ref_result_t __ovld
-intel_sub_group_avc_ref_evaluate_with_multi_reference(
-    read_only image2d_t src_image, uint packed_reference_ids,
-    uchar packed_reference_field_polarities, sampler_t vme_media_sampler,
-    intel_sub_group_avc_ref_payload_t payload);
-
-// SIC built-in functions
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_initialize(
-    ushort2 src_coord);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_configure_skc(
-    uint skip_block_partition_type, uint skip_motion_vector_mask,
-    ulong motion_vectors, uchar bidirectional_weight, uchar skip_sad_adjustment,
-    intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_configure_ipe(
-    uchar luma_intra_partition_mask, uchar intra_neighbour_availabilty,
-    uchar left_edge_luma_pixels, uchar upper_left_corner_luma_pixel,
-    uchar upper_edge_luma_pixels, uchar upper_right_edge_luma_pixels,
-    uchar intra_sad_adjustment, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_configure_ipe(
-    uchar luma_intra_partition_mask, uchar intra_neighbour_availabilty,
-    uchar left_edge_luma_pixels, uchar upper_left_corner_luma_pixel,
-    uchar upper_edge_luma_pixels, uchar upper_right_edge_luma_pixels,
-    ushort left_edge_chroma_pixels, ushort upper_left_corner_chroma_pixel,
-    ushort upper_edge_chroma_pixels, uchar intra_sad_adjustment,
-    intel_sub_group_avc_sic_payload_t payload);
-uint __ovld
-intel_sub_group_avc_sic_get_motion_vector_mask(
-    uint skip_block_partition_type, uchar direction);
-
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_intra_luma_shape_penalty(
-    uint packed_shape_cost, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_intra_luma_mode_cost_function(
-    uchar luma_mode_penalty, uint luma_packed_neighbor_modes,
-    uint luma_packed_non_dc_penalty, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_intra_chroma_mode_cost_function(
-    uchar chroma_mode_penalty, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_skc_bilinear_filter_enable(
-    intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_skc_forward_transform_enable(
-    ulong packed_sad_coefficients, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_block_based_raw_skip_sad(
-    uchar block_based_skip_type,
-    intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_sic_evaluate_ipe(
-    read_only image2d_t src_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_sic_evaluate_with_single_reference(
-    read_only image2d_t src_image, read_only image2d_t ref_image,
-    sampler_t vme_media_sampler, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_sic_evaluate_with_dual_reference(
-    read_only image2d_t src_image, read_only image2d_t fwd_ref_image,
-    read_only image2d_t bwd_ref_image, sampler_t vme_media_sampler,
-    intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_sic_evaluate_with_multi_reference(
-    read_only image2d_t src_image, uint packed_reference_ids,
-    sampler_t vme_media_sampler, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_sic_evaluate_with_multi_reference(
-    read_only image2d_t src_image, uint packed_reference_ids,
-    uchar packed_reference_field_polarities, sampler_t vme_media_sampler,
-    intel_sub_group_avc_sic_payload_t payload);
-
-uchar __ovld intel_sub_group_avc_sic_get_ipe_luma_shape(
-    intel_sub_group_avc_sic_result_t result);
-ushort __ovld intel_sub_group_avc_sic_get_best_ipe_luma_distortion(
-    intel_sub_group_avc_sic_result_t result);
-ushort __ovld intel_sub_group_avc_sic_get_best_ipe_chroma_distortion(
-    intel_sub_group_avc_sic_result_t result);
-ulong __ovld intel_sub_group_avc_sic_get_packed_ipe_luma_modes(
-    intel_sub_group_avc_sic_result_t result);
-uchar __ovld intel_sub_group_avc_sic_get_ipe_chroma_mode(
-    intel_sub_group_avc_sic_result_t result);
-uint __ovld intel_sub_group_avc_sic_get_packed_skc_luma_count_threshold(
-    intel_sub_group_avc_sic_result_t result);
-ulong __ovld intel_sub_group_avc_sic_get_packed_skc_luma_sum_threshold(
-    intel_sub_group_avc_sic_result_t result);
-ushort __ovld intel_sub_group_avc_sic_get_inter_raw_sads(
-    intel_sub_group_avc_sic_result_t result);
-
-// Wrappers
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_inter_base_multi_reference_penalty(
-    uchar reference_base_penalty, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_inter_base_multi_reference_penalty(
-    uchar reference_base_penalty, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_inter_base_multi_reference_penalty(
-    uchar reference_base_penalty, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_inter_shape_penalty(
-    ulong packed_shape_cost, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_inter_shape_penalty(
-    ulong packed_shape_cost, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_inter_shape_penalty(
-    ulong packed_shape_cost, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_inter_direction_penalty(
-    uchar direction_cost, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_inter_direction_penalty(
-    uchar direction_cost, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_inter_direction_penalty(
-    uchar direction_cost, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_motion_vector_cost_function(
-    ulong packed_cost_center_delta, uint2 packed_cost_table,
-    uchar cost_precision, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_motion_vector_cost_function(
-    ulong packed_cost_center_delta, uint2 packed_cost_table,
-    uchar cost_precision, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_motion_vector_cost_function(
-    ulong packed_cost_center_delta, uint2 packed_cost_table,
-    uchar cost_precision, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_source_interlaced_field_polarity(
-    uchar src_field_polarity, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_source_interlaced_field_polarity(
-    uchar src_field_polarity, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_source_interlaced_field_polarity(
-    uchar src_field_polarity, intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_single_reference_interlaced_field_polarity(
-    uchar ref_field_polarity, intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_single_reference_interlaced_field_polarity(
-    uchar ref_field_polarity, intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_single_reference_interlaced_field_polarity(
-    uchar ref_field_polarity, intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_dual_reference_interlaced_field_polarities(
-    uchar fwd_ref_field_polarity, uchar bwd_ref_field_polarity,
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_dual_reference_interlaced_field_polarities(
-    uchar fwd_ref_field_polarity, uchar bwd_ref_field_polarity,
-    intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_dual_reference_interlaced_field_polarities(
-    uchar fwd_ref_field_polarity, uchar bwd_ref_field_polarity,
-    intel_sub_group_avc_sic_payload_t payload);
-
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_ime_set_ac_only_haar(
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_ref_set_ac_only_haar(
-    intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_sic_set_ac_only_haar(
-    intel_sub_group_avc_sic_payload_t payload);
-
-ulong __ovld intel_sub_group_avc_ime_get_motion_vectors(
-    intel_sub_group_avc_ime_result_t result);
-ulong __ovld intel_sub_group_avc_ref_get_motion_vectors(
-    intel_sub_group_avc_ref_result_t result);
-
-ushort __ovld intel_sub_group_avc_ime_get_inter_distortions(
-    intel_sub_group_avc_ime_result_t result);
-ushort __ovld intel_sub_group_avc_ref_get_inter_distortions(
-    intel_sub_group_avc_ref_result_t result);
-ushort __ovld intel_sub_group_avc_sic_get_inter_distortions(
-    intel_sub_group_avc_sic_result_t result);
-
-ushort __ovld intel_sub_group_avc_ime_get_best_inter_distortion(
-    intel_sub_group_avc_ime_result_t result);
-ushort __ovld intel_sub_group_avc_ref_get_best_inter_distortion(
-    intel_sub_group_avc_ref_result_t result);
-
-uchar __ovld intel_sub_group_avc_ime_get_inter_major_shape(
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld intel_sub_group_avc_ref_get_inter_major_shape(
-    intel_sub_group_avc_ref_result_t result);
-uchar __ovld intel_sub_group_avc_ime_get_inter_minor_shapes(
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld intel_sub_group_avc_ref_get_inter_minor_shapes(
-    intel_sub_group_avc_ref_result_t result);
-
-uchar __ovld intel_sub_group_avc_ime_get_inter_directions(
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld intel_sub_group_avc_ref_get_inter_directions(
-    intel_sub_group_avc_ref_result_t result);
-
-uchar __ovld intel_sub_group_avc_ime_get_inter_motion_vector_count(
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld intel_sub_group_avc_ref_get_inter_motion_vector_count(
-    intel_sub_group_avc_ref_result_t result);
-
-uint __ovld intel_sub_group_avc_ime_get_inter_reference_ids(
-    intel_sub_group_avc_ime_result_t result);
-uint __ovld intel_sub_group_avc_ref_get_inter_reference_ids(
-    intel_sub_group_avc_ref_result_t result);
-
-uchar __ovld
-intel_sub_group_avc_ime_get_inter_reference_interlaced_field_polarities(
-    uint packed_reference_ids, uint packed_reference_parameter_field_polarities,
-    intel_sub_group_avc_ime_result_t result);
-uchar __ovld
-intel_sub_group_avc_ref_get_inter_reference_interlaced_field_polarities(
-    uint packed_reference_ids, uint packed_reference_parameter_field_polarities,
-    intel_sub_group_avc_ref_result_t result);
-
-// Type conversion functions
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_ime_convert_to_mce_payload(
-    intel_sub_group_avc_ime_payload_t payload);
-intel_sub_group_avc_ime_payload_t __ovld
-intel_sub_group_avc_mce_convert_to_ime_payload(
-    intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_ref_convert_to_mce_payload(
-    intel_sub_group_avc_ref_payload_t payload);
-intel_sub_group_avc_ref_payload_t __ovld
-intel_sub_group_avc_mce_convert_to_ref_payload(
-    intel_sub_group_avc_mce_payload_t payload);
-intel_sub_group_avc_mce_payload_t __ovld
-intel_sub_group_avc_sic_convert_to_mce_payload(
-    intel_sub_group_avc_sic_payload_t payload);
-intel_sub_group_avc_sic_payload_t __ovld
-intel_sub_group_avc_mce_convert_to_sic_payload(
-    intel_sub_group_avc_mce_payload_t payload);
-
-intel_sub_group_avc_mce_result_t __ovld
-intel_sub_group_avc_ime_convert_to_mce_result(
-    intel_sub_group_avc_ime_result_t result);
-intel_sub_group_avc_ime_result_t __ovld
-intel_sub_group_avc_mce_convert_to_ime_result(
-    intel_sub_group_avc_mce_result_t result);
-intel_sub_group_avc_mce_result_t __ovld
-intel_sub_group_avc_ref_convert_to_mce_result(
-    intel_sub_group_avc_ref_result_t result);
-intel_sub_group_avc_ref_result_t __ovld
-intel_sub_group_avc_mce_convert_to_ref_result(
-    intel_sub_group_avc_mce_result_t result);
-intel_sub_group_avc_mce_result_t __ovld
-intel_sub_group_avc_sic_convert_to_mce_result(
-    intel_sub_group_avc_sic_result_t result);
-intel_sub_group_avc_sic_result_t __ovld
-intel_sub_group_avc_mce_convert_to_sic_result(
-    intel_sub_group_avc_mce_result_t result);
-#pragma OPENCL EXTENSION cl_intel_device_side_avc_motion_estimation : end
-#endif // cl_intel_device_side_avc_motion_estimation
-
-#ifdef cl_amd_media_ops
-uint __ovld amd_bitalign(uint a, uint b, uint c);
-uint2 __ovld amd_bitalign(uint2 a, uint2 b, uint2 c);
-uint3 __ovld amd_bitalign(uint3 a, uint3 b, uint3 c);
-uint4 __ovld amd_bitalign(uint4 a, uint4 b, uint4 c);
-uint8 __ovld amd_bitalign(uint8 a, uint8 b, uint8 c);
-uint16 __ovld amd_bitalign(uint16 a, uint16 b, uint16 c);
-
-uint __ovld amd_bytealign(uint a, uint b, uint c);
-uint2 __ovld amd_bytealign(uint2 a, uint2 b, uint2 c);
-uint3 __ovld amd_bytealign(uint3 a, uint3 b, uint3 c);
-uint4 __ovld amd_bytealign(uint4 a, uint4 b, uint4 c);
-uint8 __ovld amd_bytealign(uint8 a, uint8 b, uint8 c);
-uint16 __ovld amd_bytealign(uint16 a, uint16 b, uint16 c);
-
-uint __ovld amd_lerp(uint a, uint b, uint c);
-uint2 __ovld amd_lerp(uint2 a, uint2 b, uint2 c);
-uint3 __ovld amd_lerp(uint3 a, uint3 b, uint3 c);
-uint4 __ovld amd_lerp(uint4 a, uint4 b, uint4 c);
-uint8 __ovld amd_lerp(uint8 a, uint8 b, uint8 c);
-uint16 __ovld amd_lerp(uint16 a, uint16 b, uint16 c);
-
-uint __ovld amd_pack(float4 v);
-
-uint __ovld amd_sad4(uint4 x, uint4 y, uint z);
-
-uint __ovld amd_sadhi(uint a, uint b, uint c);
-uint2 __ovld amd_sadhi(uint2 a, uint2 b, uint2 c);
-uint3 __ovld amd_sadhi(uint3 a, uint3 b, uint3 c);
-uint4 __ovld amd_sadhi(uint4 a, uint4 b, uint4 c);
-uint8 __ovld amd_sadhi(uint8 a, uint8 b, uint8 c);
-uint16 __ovld amd_sadhi(uint16 a, uint16 b, uint16 c);
-
-uint __ovld amd_sad(uint a, uint b, uint c);
-uint2 __ovld amd_sad(uint2 a, uint2 b, uint2 c);
-uint3 __ovld amd_sad(uint3 a, uint3 b, uint3 c);
-uint4 __ovld amd_sad(uint4 a, uint4 b, uint4 c);
-uint8 __ovld amd_sad(uint8 a, uint8 b, uint8 c);
-uint16 __ovld amd_sad(uint16 a, uint16 b, uint16 c);
-
-float __ovld amd_unpack0(uint a);
-float2 __ovld amd_unpack0(uint2 a);
-float3 __ovld amd_unpack0(uint3 a);
-float4 __ovld amd_unpack0(uint4 a);
-float8 __ovld amd_unpack0(uint8 a);
-float16 __ovld amd_unpack0(uint16 a);
-
-float __ovld amd_unpack1(uint a);
-float2 __ovld amd_unpack1(uint2 a);
-float3 __ovld amd_unpack1(uint3 a);
-float4 __ovld amd_unpack1(uint4 a);
-float8 __ovld amd_unpack1(uint8 a);
-float16 __ovld amd_unpack1(uint16 a);
-
-float __ovld amd_unpack2(uint a);
-float2 __ovld amd_unpack2(uint2 a);
-float3 __ovld amd_unpack2(uint3 a);
-float4 __ovld amd_unpack2(uint4 a);
-float8 __ovld amd_unpack2(uint8 a);
-float16 __ovld amd_unpack2(uint16 a);
-
-float __ovld amd_unpack3(uint a);
-float2 __ovld amd_unpack3(uint2 a);
-float3 __ovld amd_unpack3(uint3 a);
-float4 __ovld amd_unpack3(uint4 a);
-float8 __ovld amd_unpack3(uint8 a);
-float16 __ovld amd_unpack3(uint16 a);
-#endif // cl_amd_media_ops
-
-#ifdef cl_amd_media_ops2
-int __ovld amd_bfe(int src0, uint src1, uint src2);
-int2 __ovld amd_bfe(int2 src0, uint2 src1, uint2 src2);
-int3 __ovld amd_bfe(int3 src0, uint3 src1, uint3 src2);
-int4 __ovld amd_bfe(int4 src0, uint4 src1, uint4 src2);
-int8 __ovld amd_bfe(int8 src0, uint8 src1, uint8 src2);
-int16 __ovld amd_bfe(int16 src0, uint16 src1, uint16 src2);
-
-uint __ovld amd_bfe(uint src0, uint src1, uint src2);
-uint2 __ovld amd_bfe(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_bfe(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_bfe(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_bfe(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_bfe(uint16 src0, uint16 src1, uint16 src2);
-
-uint __ovld amd_bfm(uint src0, uint src1);
-uint2 __ovld amd_bfm(uint2 src0, uint2 src1);
-uint3 __ovld amd_bfm(uint3 src0, uint3 src1);
-uint4 __ovld amd_bfm(uint4 src0, uint4 src1);
-uint8 __ovld amd_bfm(uint8 src0, uint8 src1);
-uint16 __ovld amd_bfm(uint16 src0, uint16 src1);
-
-float __ovld amd_max3(float src0, float src1, float src2);
-float2 __ovld amd_max3(float2 src0, float2 src1, float2 src2);
-float3 __ovld amd_max3(float3 src0, float3 src1, float3 src2);
-float4 __ovld amd_max3(float4 src0, float4 src1, float4 src2);
-float8 __ovld amd_max3(float8 src0, float8 src1, float8 src2);
-float16 __ovld amd_max3(float16 src0, float16 src1, float16 src2);
-
-int __ovld amd_max3(int src0, int src1, int src2);
-int2 __ovld amd_max3(int2 src0, int2 src1, int2 src2);
-int3 __ovld amd_max3(int3 src0, int3 src1, int3 src2);
-int4 __ovld amd_max3(int4 src0, int4 src1, int4 src2);
-int8 __ovld amd_max3(int8 src0, int8 src1, int8 src2);
-int16 __ovld amd_max3(int16 src0, int16 src1, int16 src2);
-
-uint __ovld amd_max3(uint src0, uint src1, uint src2);
-uint2 __ovld amd_max3(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_max3(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_max3(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_max3(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_max3(uint16 src0, uint16 src1, uint16 src2);
-
-float __ovld amd_median3(float src0, float src1, float src2);
-float2 __ovld amd_median3(float2 src0, float2 src1, float2 src2);
-float3 __ovld amd_median3(float3 src0, float3 src1, float3 src2);
-float4 __ovld amd_median3(float4 src0, float4 src1, float4 src2);
-float8 __ovld amd_median3(float8 src0, float8 src1, float8 src2);
-float16 __ovld amd_median3(float16 src0, float16 src1, float16 src2);
-
-int __ovld amd_median3(int src0, int src1, int src2);
-int2 __ovld amd_median3(int2 src0, int2 src1, int2 src2);
-int3 __ovld amd_median3(int3 src0, int3 src1, int3 src2);
-int4 __ovld amd_median3(int4 src0, int4 src1, int4 src2);
-int8 __ovld amd_median3(int8 src0, int8 src1, int8 src2);
-int16 __ovld amd_median3(int16 src0, int16 src1, int16 src2);
-
-uint __ovld amd_median3(uint src0, uint src1, uint src2);
-uint2 __ovld amd_median3(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_median3(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_median3(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_median3(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_median3(uint16 src0, uint16 src1, uint16 src2);
-
-float __ovld amd_min3(float src0, float src1, float src);
-float2 __ovld amd_min3(float2 src0, float2 src1, float2 src);
-float3 __ovld amd_min3(float3 src0, float3 src1, float3 src);
-float4 __ovld amd_min3(float4 src0, float4 src1, float4 src);
-float8 __ovld amd_min3(float8 src0, float8 src1, float8 src);
-float16 __ovld amd_min3(float16 src0, float16 src1, float16 src);
-
-int __ovld amd_min3(int src0, int src1, int src2);
-int2 __ovld amd_min3(int2 src0, int2 src1, int2 src2);
-int3 __ovld amd_min3(int3 src0, int3 src1, int3 src2);
-int4 __ovld amd_min3(int4 src0, int4 src1, int4 src2);
-int8 __ovld amd_min3(int8 src0, int8 src1, int8 src2);
-int16 __ovld amd_min3(int16 src0, int16 src1, int16 src2);
-
-uint __ovld amd_min3(uint src0, uint src1, uint src2);
-uint2 __ovld amd_min3(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_min3(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_min3(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_min3(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_min3(uint16 src0, uint16 src1, uint16 src2);
-
-ulong __ovld amd_mqsad(ulong src0, uint src1, ulong src2);
-ulong2 __ovld amd_mqsad(ulong2 src0, uint2 src1, ulong2 src2);
-ulong3 __ovld amd_mqsad(ulong3 src0, uint3 src1, ulong3 src2);
-ulong4 __ovld amd_mqsad(ulong4 src0, uint4 src1, ulong4 src2);
-ulong8 __ovld amd_mqsad(ulong8 src0, uint8 src1, ulong8 src2);
-ulong16 __ovld amd_mqsad(ulong16 src0, uint16 src1, ulong16 src2);
-
-ulong __ovld amd_qsad(ulong src0, uint src1, ulong src2);
-ulong2 __ovld amd_qsad(ulong2 src0, uint2 src1, ulong2 src2);
-ulong3 __ovld amd_qsad(ulong3 src0, uint3 src1, ulong3 src2);
-ulong4 __ovld amd_qsad(ulong4 src0, uint4 src1, ulong4 src2);
-ulong8 __ovld amd_qsad(ulong8 src0, uint8 src1, ulong8 src2);
-ulong16 __ovld amd_qsad(ulong16 src0, uint16 src1, ulong16 src2);
-
-uint __ovld amd_msad(uint src0, uint src1, uint src2);
-uint2 __ovld amd_msad(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_msad(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_msad(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_msad(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_msad(uint16 src0, uint16 src1, uint16 src2);
-
-uint __ovld amd_sadd(uint src0, uint src1, uint src2);
-uint2 __ovld amd_sadd(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_sadd(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_sadd(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_sadd(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_sadd(uint16 src0, uint16 src1, uint16 src2);
-
-uint __ovld amd_sadw(uint src0, uint src1, uint src2);
-uint2 __ovld amd_sadw(uint2 src0, uint2 src1, uint2 src2);
-uint3 __ovld amd_sadw(uint3 src0, uint3 src1, uint3 src2);
-uint4 __ovld amd_sadw(uint4 src0, uint4 src1, uint4 src2);
-uint8 __ovld amd_sadw(uint8 src0, uint8 src1, uint8 src2);
-uint16 __ovld amd_sadw(uint16 src0, uint16 src1, uint16 src2);
-#endif // cl_amd_media_ops2
-
-#if defined(cl_arm_integer_dot_product_int8)
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_int8 : begin
-uint __ovld arm_dot(uchar4 a, uchar4 b);
-int __ovld arm_dot(char4 a, char4 b);
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_int8 : end
-#endif // defined(cl_arm_integer_dot_product_int8)
-
-#if defined(cl_arm_integer_dot_product_accumulate_int8)
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_int8 : begin
-uint __ovld arm_dot_acc(uchar4 a, uchar4 b, uint c);
-int __ovld arm_dot_acc(char4 a, char4 b, int c);
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_int8 : end
-#endif // defined(cl_arm_integer_dot_product_accumulate_int8)
-
-#if defined(cl_arm_integer_dot_product_accumulate_int16)
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_int16 : begin
-uint __ovld arm_dot_acc(ushort2 a, ushort2 b, uint c);
-int __ovld arm_dot_acc(short2 a, short2 b, int c);
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_int16 : end
-#endif // defined(cl_arm_integer_dot_product_accumulate_int16)
-
-#if defined(cl_arm_integer_dot_product_accumulate_saturate_int8)
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_saturate_int8 : begin
-uint __ovld arm_dot_acc_sat(uchar4 a, uchar4 b, uint c);
-int __ovld arm_dot_acc_sat(char4 a, char4 b, int c);
-#pragma OPENCL EXTENSION cl_arm_integer_dot_product_accumulate_saturate_int8 : end
-#endif // defined(cl_arm_integer_dot_product_accumulate_saturate_int8)
-
-// Disable any extensions we may have enabled previously.
-#pragma OPENCL EXTENSION all : disable
-
-#undef __cnfn
-#undef __ovld
-#endif //_OPENCL_H_
diff --git a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h b/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
deleted file mode 100644
index 617ce24..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/emmintrin.h
+++ /dev/null
@@ -1,2318 +0,0 @@
-/*===---- emmintrin.h - Implementation of SSE2 intrinsics on PowerPC -------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-/* Implemented from the specification included in the Intel C++ Compiler
-   User Guide and Reference, version 9.0.  */
-
-#ifndef NO_WARN_X86_INTRINSICS
-/* This header file is to help porting code using Intel intrinsics
-   explicitly from x86_64 to powerpc64/powerpc64le.
-
-   Since X86 SSE2 intrinsics mainly handles __m128i and __m128d type,
-   PowerPC VMX/VSX ISA is a good match for vector float SIMD operations.
-   However scalar float operations in vector (XMM) registers require
-   the POWER8 VSX ISA (2.07) level. There are differences for data
-   format and placement of float scalars in the vector register, which
-   require extra steps to match SSE2 scalar float semantics on POWER.
-
-   It should be noted that there's much difference between X86_64's
-   MXSCR and PowerISA's FPSCR/VSCR registers. It's recommended to use
-   portable <fenv.h> instead of access MXSCR directly.
-
-   Most SSE2 scalar float intrinsic operations can be performed more
-   efficiently as C language float scalar operations or optimized to
-   use vector SIMD operations. We recommend this for new applications.
-*/
-#error "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this error."
-#endif
-
-#ifndef EMMINTRIN_H_
-#define EMMINTRIN_H_
-
-#include <altivec.h>
-
-/* We need definitions from the SSE header files.  */
-#include <xmmintrin.h>
-
-/* SSE2 */
-typedef __vector double __v2df;
-typedef __vector long long __v2di;
-typedef __vector unsigned long long __v2du;
-typedef __vector int __v4si;
-typedef __vector unsigned int __v4su;
-typedef __vector short __v8hi;
-typedef __vector unsigned short __v8hu;
-typedef __vector signed char __v16qi;
-typedef __vector unsigned char __v16qu;
-
-/* The Intel API is flexible enough that we must allow aliasing with other
-   vector types, and their scalar components.  */
-typedef long long __m128i __attribute__ ((__vector_size__ (16), __may_alias__));
-typedef double __m128d __attribute__ ((__vector_size__ (16), __may_alias__));
-
-/* Unaligned version of the same types.  */
-typedef long long __m128i_u __attribute__ ((__vector_size__ (16), __may_alias__, __aligned__ (1)));
-typedef double __m128d_u __attribute__ ((__vector_size__ (16), __may_alias__, __aligned__ (1)));
-
-/* Define two value permute mask.  */
-#define _MM_SHUFFLE2(x,y) (((x) << 1) | (y))
-
-/* Create a vector with element 0 as F and the rest zero.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_sd (double __F)
-{
-  return __extension__ (__m128d){ __F, 0.0 };
-}
-
-/* Create a vector with both elements equal to F.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_pd (double __F)
-{
-  return __extension__ (__m128d){ __F, __F };
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_pd1 (double __F)
-{
-  return _mm_set1_pd (__F);
-}
-
-/* Create a vector with the lower value X and upper value W.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_pd (double __W, double __X)
-{
-  return __extension__ (__m128d){ __X, __W };
-}
-
-/* Create a vector with the lower value W and upper value X.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_pd (double __W, double __X)
-{
-  return __extension__ (__m128d){ __W, __X };
-}
-
-/* Create an undefined vector.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_undefined_pd (void)
-{
-  __m128d __Y = __Y;
-  return __Y;
-}
-
-/* Create a vector of zeros.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setzero_pd (void)
-{
-  return (__m128d) vec_splats (0);
-}
-
-/* Sets the low DPFP value of A from the low value of B.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_move_sd (__m128d __A, __m128d __B)
-{
-  __v2df result = (__v2df) __A;
-  result [0] = ((__v2df) __B)[0];
-  return (__m128d) result;
-}
-
-/* Load two DPFP values from P.  The address must be 16-byte aligned.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_pd (double const *__P)
-{
-  return ((__m128d)vec_ld(0, (__v16qu*)__P));
-}
-
-/* Load two DPFP values from P.  The address need not be 16-byte aligned.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadu_pd (double const *__P)
-{
-  return (vec_vsx_ld(0, __P));
-}
-
-/* Create a vector with all two elements equal to *P.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load1_pd (double const *__P)
-{
-  return (vec_splats (*__P));
-}
-
-/* Create a vector with element 0 as *P and the rest zero.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_sd (double const *__P)
-{
-  return _mm_set_sd (*__P);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_pd1 (double const *__P)
-{
-  return _mm_load1_pd (__P);
-}
-
-/* Load two DPFP values in reverse order.  The address must be aligned.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadr_pd (double const *__P)
-{
-  __v2df __tmp = _mm_load_pd (__P);
-  return (__m128d)vec_xxpermdi (__tmp, __tmp, 2);
-}
-
-/* Store two DPFP values.  The address must be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_pd (double *__P, __m128d __A)
-{
-  vec_st((__v16qu)__A, 0, (__v16qu*)__P);
-}
-
-/* Store two DPFP values.  The address need not be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storeu_pd (double *__P, __m128d __A)
-{
-  *(__m128d_u *)__P = __A;
-}
-
-/* Stores the lower DPFP value.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_sd (double *__P, __m128d __A)
-{
-  *__P = ((__v2df)__A)[0];
-}
-
-extern __inline double __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsd_f64 (__m128d __A)
-{
-  return ((__v2df)__A)[0];
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storel_pd (double *__P, __m128d __A)
-{
-  _mm_store_sd (__P, __A);
-}
-
-/* Stores the upper DPFP value.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storeh_pd (double *__P, __m128d __A)
-{
-  *__P = ((__v2df)__A)[1];
-}
-/* Store the lower DPFP value across two words.
-   The address must be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store1_pd (double *__P, __m128d __A)
-{
-  _mm_store_pd (__P, vec_splat (__A, 0));
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_pd1 (double *__P, __m128d __A)
-{
-  _mm_store1_pd (__P, __A);
-}
-
-/* Store two DPFP values in reverse order.  The address must be aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storer_pd (double *__P, __m128d __A)
-{
-  _mm_store_pd (__P, vec_xxpermdi (__A, __A, 2));
-}
-
-/* Intel intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi128_si64 (__m128i __A)
-{
-  return ((__v2di)__A)[0];
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi128_si64x (__m128i __A)
-{
-  return ((__v2di)__A)[0];
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) ((__v2df)__A + (__v2df)__B);
-}
-
-/* Add the lower double-precision (64-bit) floating-point element in
-   a and b, store the result in the lower element of dst, and copy
-   the upper element from a to the upper element of dst. */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_sd (__m128d __A, __m128d __B)
-{
-  __A[0] = __A[0] + __B[0];
-  return (__A);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) ((__v2df)__A - (__v2df)__B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_sd (__m128d __A, __m128d __B)
-{
-  __A[0] = __A[0] - __B[0];
-  return (__A);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) ((__v2df)__A * (__v2df)__B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_sd (__m128d __A, __m128d __B)
-{
-  __A[0] = __A[0] * __B[0];
-  return (__A);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_div_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) ((__v2df)__A / (__v2df)__B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_div_sd (__m128d __A, __m128d __B)
-{
-  __A[0] = __A[0] / __B[0];
-  return (__A);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sqrt_pd (__m128d __A)
-{
-  return (vec_sqrt (__A));
-}
-
-/* Return pair {sqrt (B[0]), A[1]}.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sqrt_sd (__m128d __A, __m128d __B)
-{
-  __v2df c;
-  c = vec_sqrt ((__v2df) _mm_set1_pd (__B[0]));
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_pd (__m128d __A, __m128d __B)
-{
-  return (vec_min (__A, __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = vec_min (a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_pd (__m128d __A, __m128d __B)
-{
-  return (vec_max (__A, __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = vec_max (a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmpeq ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmplt ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmple_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmple ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmpgt ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpge_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmpge ((__v2df) __A,(__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpneq_pd (__m128d __A, __m128d __B)
-{
-  __v2df temp = (__v2df) vec_cmpeq ((__v2df) __A, (__v2df)__B);
-  return ((__m128d)vec_nor (temp, temp));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnlt_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmpge ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnle_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmpgt ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpngt_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmple ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnge_pd (__m128d __A, __m128d __B)
-{
-  return ((__m128d)vec_cmplt ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpord_pd (__m128d __A, __m128d __B)
-{
-#if _ARCH_PWR8
-  __v2du c, d;
-  /* Compare against self will return false (0's) if NAN.  */
-  c = (__v2du)vec_cmpeq (__A, __A);
-  d = (__v2du)vec_cmpeq (__B, __B);
-#else
-  __v2du a, b;
-  __v2du c, d;
-  const __v2du double_exp_mask  = {0x7ff0000000000000, 0x7ff0000000000000};
-  a = (__v2du)vec_abs ((__v2df)__A);
-  b = (__v2du)vec_abs ((__v2df)__B);
-  c = (__v2du)vec_cmpgt (double_exp_mask, a);
-  d = (__v2du)vec_cmpgt (double_exp_mask, b);
-#endif
-  /* A != NAN and B != NAN.  */
-  return ((__m128d)vec_and(c, d));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpunord_pd (__m128d __A, __m128d __B)
-{
-#if _ARCH_PWR8
-  __v2du c, d;
-  /* Compare against self will return false (0's) if NAN.  */
-  c = (__v2du)vec_cmpeq ((__v2df)__A, (__v2df)__A);
-  d = (__v2du)vec_cmpeq ((__v2df)__B, (__v2df)__B);
-  /* A == NAN OR B == NAN converts too:
-     NOT(A != NAN) OR NOT(B != NAN).  */
-  c = vec_nor (c, c);
-  return ((__m128d)vec_orc(c, d));
-#else
-  __v2du c, d;
-  /* Compare against self will return false (0's) if NAN.  */
-  c = (__v2du)vec_cmpeq ((__v2df)__A, (__v2df)__A);
-  d = (__v2du)vec_cmpeq ((__v2df)__B, (__v2df)__B);
-  /* Convert the true ('1's) is NAN.  */
-  c = vec_nor (c, c);
-  d = vec_nor (d, d);
-  return ((__m128d)vec_or(c, d));
-#endif
-}
-
-extern __inline  __m128d  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_sd(__m128d  __A, __m128d  __B)
-{
-  __v2df a, b, c;
-  /* PowerISA VSX does not allow partial (for just lower double)
-     results. So to insure we don't generate spurious exceptions
-     (from the upper double values) we splat the lower double
-     before we do the operation. */
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmpeq(a, b);
-  /* Then we merge the lower double result with the original upper
-     double from __A.  */
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmplt(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmple_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmple(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmpgt(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpge_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmpge(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpneq_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  c = (__v2df) vec_cmpeq(a, b);
-  c = vec_nor (c, c);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnlt_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  /* Not less than is just greater than or equal.  */
-  c = (__v2df) vec_cmpge(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnle_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  /* Not less than or equal is just greater than.  */
-  c = (__v2df) vec_cmpge(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpngt_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  /* Not greater than is just less than or equal.  */
-  c = (__v2df) vec_cmple(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnge_sd (__m128d __A, __m128d __B)
-{
-  __v2df a, b, c;
-  a = vec_splats (__A[0]);
-  b = vec_splats (__B[0]);
-  /* Not greater than or equal is just less than.  */
-  c = (__v2df) vec_cmplt(a, b);
-  return (__m128d) _mm_setr_pd (c[0], __A[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpord_sd (__m128d __A, __m128d __B)
-{
-  __v2df r;
-  r = (__v2df)_mm_cmpord_pd (vec_splats (__A[0]), vec_splats (__B[0]));
-  return (__m128d) _mm_setr_pd (r[0], ((__v2df)__A)[1]);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpunord_sd (__m128d __A, __m128d __B)
-{
-  __v2df r;
-  r = _mm_cmpunord_pd (vec_splats (__A[0]), vec_splats (__B[0]));
-  return (__m128d) _mm_setr_pd (r[0], __A[1]);
-}
-
-/* FIXME
-   The __mm_comi??_sd and __mm_ucomi??_sd implementations below are
-   exactly the same because GCC for PowerPC only generates unordered
-   compares (scalar and vector).
-   Technically __mm_comieq_sp et all should be using the ordered
-   compare and signal for QNaNs.  The __mm_ucomieq_sd et all should
-   be OK.   */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comieq_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] == __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comilt_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] < __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comile_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] <= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comigt_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] > __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comige_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] >= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comineq_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] != __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomieq_sd (__m128d __A, __m128d __B)
-{
-	return (__A[0] == __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomilt_sd (__m128d __A, __m128d __B)
-{
-	return (__A[0] < __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomile_sd (__m128d __A, __m128d __B)
-{
-	return (__A[0] <= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomigt_sd (__m128d __A, __m128d __B)
-{
-	return (__A[0] > __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomige_sd (__m128d __A, __m128d __B)
-{
-	return (__A[0] >= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomineq_sd (__m128d __A, __m128d __B)
-{
-  return (__A[0] != __B[0]);
-}
-
-/* Create a vector of Qi, where i is the element number.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_epi64x (long long __q1, long long __q0)
-{
-  return __extension__ (__m128i)(__v2di){ __q0, __q1 };
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_epi64 (__m64 __q1,  __m64 __q0)
-{
-  return _mm_set_epi64x ((long long)__q1, (long long)__q0);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_epi32 (int __q3, int __q2, int __q1, int __q0)
-{
-  return __extension__ (__m128i)(__v4si){ __q0, __q1, __q2, __q3 };
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_epi16 (short __q7, short __q6, short __q5, short __q4,
-	       short __q3, short __q2, short __q1, short __q0)
-{
-  return __extension__ (__m128i)(__v8hi){
-    __q0, __q1, __q2, __q3, __q4, __q5, __q6, __q7 };
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_epi8 (char __q15, char __q14, char __q13, char __q12,
-	      char __q11, char __q10, char __q09, char __q08,
-	      char __q07, char __q06, char __q05, char __q04,
-	      char __q03, char __q02, char __q01, char __q00)
-{
-  return __extension__ (__m128i)(__v16qi){
-    __q00, __q01, __q02, __q03, __q04, __q05, __q06, __q07,
-    __q08, __q09, __q10, __q11, __q12, __q13, __q14, __q15
-  };
-}
-
-/* Set all of the elements of the vector to A.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_epi64x (long long __A)
-{
-  return _mm_set_epi64x (__A, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_epi64 (__m64 __A)
-{
-  return _mm_set_epi64 (__A, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_epi32 (int __A)
-{
-  return _mm_set_epi32 (__A, __A, __A, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_epi16 (short __A)
-{
-  return _mm_set_epi16 (__A, __A, __A, __A, __A, __A, __A, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_epi8 (char __A)
-{
-  return _mm_set_epi8 (__A, __A, __A, __A, __A, __A, __A, __A,
-		       __A, __A, __A, __A, __A, __A, __A, __A);
-}
-
-/* Create a vector of Qi, where i is the element number.
-   The parameter order is reversed from the _mm_set_epi* functions.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_epi64 (__m64 __q0, __m64 __q1)
-{
-  return _mm_set_epi64 (__q1, __q0);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_epi32 (int __q0, int __q1, int __q2, int __q3)
-{
-  return _mm_set_epi32 (__q3, __q2, __q1, __q0);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_epi16 (short __q0, short __q1, short __q2, short __q3,
-	        short __q4, short __q5, short __q6, short __q7)
-{
-  return _mm_set_epi16 (__q7, __q6, __q5, __q4, __q3, __q2, __q1, __q0);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_epi8 (char __q00, char __q01, char __q02, char __q03,
-	       char __q04, char __q05, char __q06, char __q07,
-	       char __q08, char __q09, char __q10, char __q11,
-	       char __q12, char __q13, char __q14, char __q15)
-{
-  return _mm_set_epi8 (__q15, __q14, __q13, __q12, __q11, __q10, __q09, __q08,
-		       __q07, __q06, __q05, __q04, __q03, __q02, __q01, __q00);
-}
-
-/* Create a vector with element 0 as *P and the rest zero.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_si128 (__m128i const *__P)
-{
-  return *__P;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadu_si128 (__m128i_u const *__P)
-{
-  return (__m128i) (vec_vsx_ld(0, (signed int const *)__P));
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadl_epi64 (__m128i_u const *__P)
-{
-  return _mm_set_epi64 ((__m64)0LL, *(__m64 *)__P);
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_si128 (__m128i *__P, __m128i __B)
-{
-  vec_st ((__v16qu) __B, 0, (__v16qu*)__P);
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storeu_si128 (__m128i_u *__P, __m128i __B)
-{
-  *__P = __B;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storel_epi64 (__m128i_u *__P, __m128i __B)
-{
-  *(long long *)__P = ((__v2di)__B)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movepi64_pi64 (__m128i_u __B)
-{
-  return (__m64) ((__v2di)__B)[0];
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movpi64_epi64 (__m64 __A)
-{
-  return _mm_set_epi64 ((__m64)0LL, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_move_epi64 (__m128i __A)
-{
-  return _mm_set_epi64 ((__m64)0LL, (__m64)__A[0]);
-}
-
-/* Create an undefined vector.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_undefined_si128 (void)
-{
-  __m128i __Y = __Y;
-  return __Y;
-}
-
-/* Create a vector of zeros.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setzero_si128 (void)
-{
-  return __extension__ (__m128i)(__v4si){ 0, 0, 0, 0 };
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtepi32_pd (__m128i __A)
-{
-  __v2di val;
-  /* For LE need to generate Vector Unpack Low Signed Word.
-     Which is generated from unpackh.  */
-  val = (__v2di)vec_unpackh ((__v4si)__A);
-
-  return (__m128d)vec_ctf (val, 0);
-}
-#endif
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtepi32_ps (__m128i __A)
-{
-  return ((__m128)vec_ctf((__v4si)__A, 0));
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpd_epi32 (__m128d __A)
-{
-  __v2df rounded = vec_rint (__A);
-  __v4si result, temp;
-  const __v4si vzero =
-    { 0, 0, 0, 0 };
-
-  /* VSX Vector truncate Double-Precision to integer and Convert to
-   Signed Integer Word format with Saturate.  */
-  __asm__(
-      "xvcvdpsxws %x0,%x1"
-      : "=wa" (temp)
-      : "wa" (rounded)
-      : );
-
-#ifdef _ARCH_PWR8
-  temp = vec_mergeo (temp, temp);
-  result = (__v4si) vec_vpkudum ((__vector long long) temp,
-				 (__vector long long) vzero);
-#else
-  {
-    const __v16qu pkperm = {0x00, 0x01, 0x02, 0x03, 0x08, 0x09, 0x0a, 0x0b,
-	0x14, 0x15, 0x16, 0x17, 0x1c, 0x1d, 0x1e, 0x1f };
-    result = (__v4si) vec_perm ((__v16qu) temp, (__v16qu) vzero, pkperm);
-  }
-#endif
-  return (__m128i) result;
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpd_pi32 (__m128d __A)
-{
-  __m128i result = _mm_cvtpd_epi32(__A);
-
-  return (__m64) result[0];
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpd_ps (__m128d __A)
-{
-  __v4sf result;
-  __v4si temp;
-  const __v4si vzero = { 0, 0, 0, 0 };
-
-  __asm__(
-      "xvcvdpsp %x0,%x1"
-      : "=wa" (temp)
-      : "wa" (__A)
-      : );
-
-#ifdef _ARCH_PWR8
-  temp = vec_mergeo (temp, temp);
-  result = (__v4sf) vec_vpkudum ((__vector long long) temp,
-				 (__vector long long) vzero);
-#else
-  {
-    const __v16qu pkperm = {0x00, 0x01, 0x02, 0x03, 0x08, 0x09, 0x0a, 0x0b,
-	0x14, 0x15, 0x16, 0x17, 0x1c, 0x1d, 0x1e, 0x1f };
-    result = (__v4sf) vec_perm ((__v16qu) temp, (__v16qu) vzero, pkperm);
-  }
-#endif
-  return ((__m128)result);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttpd_epi32 (__m128d __A)
-{
-  __v4si result;
-  __v4si temp;
-  const __v4si vzero = { 0, 0, 0, 0 };
-
-  /* VSX Vector truncate Double-Precision to integer and Convert to
-   Signed Integer Word format with Saturate.  */
-  __asm__(
-      "xvcvdpsxws %x0,%x1"
-      : "=wa" (temp)
-      : "wa" (__A)
-      : );
-
-#ifdef _ARCH_PWR8
-  temp = vec_mergeo (temp, temp);
-  result = (__v4si) vec_vpkudum ((__vector long long) temp,
-				 (__vector long long) vzero);
-#else
-  {
-    const __v16qu pkperm = {0x00, 0x01, 0x02, 0x03, 0x08, 0x09, 0x0a, 0x0b,
-	0x14, 0x15, 0x16, 0x17, 0x1c, 0x1d, 0x1e, 0x1f };
-    result = (__v4si) vec_perm ((__v16qu) temp, (__v16qu) vzero, pkperm);
-  }
-#endif
-
-  return ((__m128i) result);
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttpd_pi32 (__m128d __A)
-{
-  __m128i result = _mm_cvttpd_epi32 (__A);
-
-  return (__m64) result[0];
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi128_si32 (__m128i __A)
-{
-  return ((__v4si)__A)[0];
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpi32_pd (__m64 __A)
-{
-  __v4si temp;
-  __v2di tmp2;
-  __v2df result;
-
-  temp = (__v4si)vec_splats (__A);
-  tmp2 = (__v2di)vec_unpackl (temp);
-  result = vec_ctf ((__vector signed long long) tmp2, 0);
-  return (__m128d)result;
-}
-#endif
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtps_epi32 (__m128 __A)
-{
-  __v4sf rounded;
-  __v4si result;
-
-  rounded = vec_rint((__v4sf) __A);
-  result = vec_cts (rounded, 0);
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttps_epi32 (__m128 __A)
-{
-  __v4si result;
-
-  result = vec_cts ((__v4sf) __A, 0);
-  return (__m128i) result;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtps_pd (__m128 __A)
-{
-  /* Check if vec_doubleh is defined by <altivec.h>. If so use that. */
-#ifdef vec_doubleh
-  return (__m128d) vec_doubleh ((__v4sf)__A);
-#else
-  /* Otherwise the compiler is not current and so need to generate the
-     equivalent code.  */
-  __v4sf a = (__v4sf)__A;
-  __v4sf temp;
-  __v2df result;
-#ifdef __LITTLE_ENDIAN__
-  /* The input float values are in elements {[0], [1]} but the convert
-     instruction needs them in elements {[1], [3]}, So we use two
-     shift left double vector word immediates to get the elements
-     lined up.  */
-  temp = __builtin_vsx_xxsldwi (a, a, 3);
-  temp = __builtin_vsx_xxsldwi (a, temp, 2);
-#else
-  /* The input float values are in elements {[0], [1]} but the convert
-     instruction needs them in elements {[0], [2]}, So we use two
-     shift left double vector word immediates to get the elements
-     lined up.  */
-  temp = vec_vmrghw (a, a);
-#endif
-  __asm__(
-      " xvcvspdp %x0,%x1"
-      : "=wa" (result)
-      : "wa" (temp)
-      : );
-  return (__m128d) result;
-#endif
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsd_si32 (__m128d __A)
-{
-  __v2df rounded = vec_rint((__v2df) __A);
-  int result = ((__v2df)rounded)[0];
-
-  return result;
-}
-/* Intel intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsd_si64 (__m128d __A)
-{
-  __v2df rounded = vec_rint ((__v2df) __A );
-  long long result = ((__v2df) rounded)[0];
-
-  return result;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsd_si64x (__m128d __A)
-{
-  return _mm_cvtsd_si64 ((__v2df)__A);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttsd_si32 (__m128d __A)
-{
-  int result = ((__v2df)__A)[0];
-
-  return result;
-}
-
-/* Intel intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttsd_si64 (__m128d __A)
-{
-  long long result = ((__v2df)__A)[0];
-
-  return result;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttsd_si64x (__m128d __A)
-{
-  return _mm_cvttsd_si64 (__A);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsd_ss (__m128 __A, __m128d __B)
-{
-  __v4sf result = (__v4sf)__A;
-
-#ifdef __LITTLE_ENDIAN__
-  __v4sf temp_s;
-  /* Copy double element[0] to element [1] for conversion.  */
-  __v2df temp_b = vec_splat((__v2df)__B, 0);
-
-  /* Pre-rotate __A left 3 (logically right 1) elements.  */
-  result = __builtin_vsx_xxsldwi (result, result, 3);
-  /* Convert double to single float scalar in a vector.  */
-  __asm__(
-      "xscvdpsp %x0,%x1"
-      : "=wa" (temp_s)
-      : "wa" (temp_b)
-      : );
-  /* Shift the resulting scalar into vector element [0].  */
-  result = __builtin_vsx_xxsldwi (result, temp_s, 1);
-#else
-  result [0] = ((__v2df)__B)[0];
-#endif
-  return (__m128) result;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi32_sd (__m128d __A, int __B)
-{
-  __v2df result = (__v2df)__A;
-  double db = __B;
-  result [0] = db;
-  return (__m128d)result;
-}
-
-/* Intel intrinsic.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64_sd (__m128d __A, long long __B)
-{
-  __v2df result = (__v2df)__A;
-  double db = __B;
-  result [0] = db;
-  return (__m128d)result;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64x_sd (__m128d __A, long long __B)
-{
-  return _mm_cvtsi64_sd (__A, __B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtss_sd (__m128d __A, __m128 __B)
-{
-#ifdef __LITTLE_ENDIAN__
-  /* Use splat to move element [0] into position for the convert. */
-  __v4sf temp = vec_splat ((__v4sf)__B, 0);
-  __v2df res;
-  /* Convert single float scalar to double in a vector.  */
-  __asm__(
-      "xscvspdp %x0,%x1"
-      : "=wa" (res)
-      : "wa" (temp)
-      : );
-  return (__m128d) vec_mergel (res, (__v2df)__A);
-#else
-  __v2df res = (__v2df)__A;
-  res [0] = ((__v4sf)__B) [0];
-  return (__m128d) res;
-#endif
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_shuffle_pd(__m128d __A, __m128d __B, const int __mask)
-{
-  __vector double result;
-  const int litmsk = __mask & 0x3;
-
-  if (litmsk == 0)
-    result = vec_mergeh (__A, __B);
-#if __GNUC__ < 6
-  else if (litmsk == 1)
-    result = vec_xxpermdi (__B, __A, 2);
-  else if (litmsk == 2)
-    result = vec_xxpermdi (__B, __A, 1);
-#else
-  else if (litmsk == 1)
-    result = vec_xxpermdi (__A, __B, 2);
-  else if (litmsk == 2)
-    result = vec_xxpermdi (__A, __B, 1);
-#endif
-  else
-    result = vec_mergel (__A, __B);
-
-  return result;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) vec_mergel ((__v2df)__A, (__v2df)__B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_pd (__m128d __A, __m128d __B)
-{
-  return (__m128d) vec_mergeh ((__v2df)__A, (__v2df)__B);
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadh_pd (__m128d __A, double const *__B)
-{
-  __v2df result = (__v2df)__A;
-  result [1] = *__B;
-  return (__m128d)result;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadl_pd (__m128d __A, double const *__B)
-{
-  __v2df result = (__v2df)__A;
-  result [0] = *__B;
-  return (__m128d)result;
-}
-
-#ifdef _ARCH_PWR8
-/* Intrinsic functions that require PowerISA 2.07 minimum.  */
-
-/* Creates a 2-bit mask from the most significant bits of the DPFP values.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movemask_pd (__m128d  __A)
-{
-  __vector unsigned long long result;
-  static const __vector unsigned int perm_mask =
-    {
-#ifdef __LITTLE_ENDIAN__
-	0x80800040, 0x80808080, 0x80808080, 0x80808080
-#else
-      0x80808080, 0x80808080, 0x80808080, 0x80804000
-#endif
-    };
-
-  result = ((__vector unsigned long long)
-	    vec_vbpermq ((__vector unsigned char) __A,
-			 (__vector unsigned char) perm_mask));
-
-#ifdef __LITTLE_ENDIAN__
-  return result[1];
-#else
-  return result[0];
-#endif
-}
-#endif /* _ARCH_PWR8 */
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_packs_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_packs ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_packs_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_packs ((__v4si)__A, (__v4si)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_packus_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_packsu ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergel ((__v16qu)__A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergel ((__v8hu)__A, (__v8hu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergel ((__v4su)__A, (__v4su)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_epi64 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergel ((__vector long long) __A,
-			       (__vector long long) __B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergeh ((__v16qu)__A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergeh ((__v8hi)__A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergeh ((__v4si)__A, (__v4si)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_epi64 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_mergeh ((__vector long long) __A,
-			       (__vector long long) __B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v16qu)__A + (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v8hu)__A + (__v8hu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v4su)__A + (__v4su)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_epi64 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v2du)__A + (__v2du)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_adds_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_adds ((__v16qi)__A, (__v16qi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_adds_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_adds ((__v8hi)__A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_adds_epu8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_adds ((__v16qu)__A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_adds_epu16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_adds ((__v8hu)__A, (__v8hu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v16qu)__A - (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v8hu)__A - (__v8hu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v4su)__A - (__v4su)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_epi64 (__m128i __A, __m128i __B)
-{
-  return (__m128i) ((__v2du)__A - (__v2du)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_subs_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_subs ((__v16qi)__A, (__v16qi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_subs_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_subs ((__v8hi)__A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_subs_epu8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_subs ((__v16qu)__A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_subs_epu16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_subs ((__v8hu)__A, (__v8hu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_madd_epi16 (__m128i __A, __m128i __B)
-{
-  __vector signed int zero = {0, 0, 0, 0};
-
-  return (__m128i) vec_vmsumshm ((__v8hi)__A, (__v8hi)__B, zero);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mulhi_epi16 (__m128i __A, __m128i __B)
-{
-  __vector signed int w0, w1;
-
-  __vector unsigned char xform1 = {
-#ifdef __LITTLE_ENDIAN__
-      0x02, 0x03, 0x12, 0x13,  0x06, 0x07, 0x16, 0x17,
-      0x0A, 0x0B, 0x1A, 0x1B,  0x0E, 0x0F, 0x1E, 0x1F
-#else
-      0x00, 0x01, 0x10, 0x11,  0x04, 0x05, 0x14, 0x15,
-      0x08, 0x09, 0x18, 0x19,  0x0C, 0x0D, 0x1C, 0x1D
-#endif
-    };
-
-  w0 = vec_vmulesh ((__v8hi)__A, (__v8hi)__B);
-  w1 = vec_vmulosh ((__v8hi)__A, (__v8hi)__B);
-  return (__m128i) vec_perm (w0, w1, xform1);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mullo_epi16 (__m128i __A, __m128i __B)
-{
-    return (__m128i) ((__v8hi)__A * (__v8hi)__B);
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_su32 (__m64 __A, __m64 __B)
-{
-  unsigned int a = __A;
-  unsigned int b = __B;
-
-  return ((__m64)a * (__m64)b);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_epu32 (__m128i __A, __m128i __B)
-{
-#if __GNUC__ < 8
-  __v2du result;
-
-#ifdef __LITTLE_ENDIAN__
-  /* VMX Vector Multiply Odd Unsigned Word.  */
-  __asm__(
-      "vmulouw %0,%1,%2"
-      : "=v" (result)
-      : "v" (__A), "v" (__B)
-      : );
-#else
-  /* VMX Vector Multiply Even Unsigned Word.  */
-  __asm__(
-      "vmuleuw %0,%1,%2"
-      : "=v" (result)
-      : "v" (__A), "v" (__B)
-      : );
-#endif
-  return (__m128i) result;
-#else
-  return (__m128i) vec_mule ((__v4su)__A, (__v4su)__B);
-#endif
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_slli_epi16 (__m128i __A, int __B)
-{
-  __v8hu lshift;
-  __v8hi result = { 0, 0, 0, 0, 0, 0, 0, 0 };
-
-  if (__B >= 0 && __B < 16)
-    {
-      if (__builtin_constant_p(__B))
-	lshift = (__v8hu) vec_splat_s16(__B);
-      else
-	lshift = vec_splats ((unsigned short) __B);
-
-      result = vec_sl ((__v8hi) __A, lshift);
-    }
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_slli_epi32 (__m128i __A, int __B)
-{
-  __v4su lshift;
-  __v4si result = { 0, 0, 0, 0 };
-
-  if (__B >= 0 && __B < 32)
-    {
-      if (__builtin_constant_p(__B) && __B < 16)
-	lshift = (__v4su) vec_splat_s32(__B);
-      else
-	lshift = vec_splats ((unsigned int) __B);
-
-      result = vec_sl ((__v4si) __A, lshift);
-    }
-
-  return (__m128i) result;
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_slli_epi64 (__m128i __A, int __B)
-{
-  __v2du lshift;
-  __v2di result = { 0, 0 };
-
-  if (__B >= 0 && __B < 64)
-    {
-      if (__builtin_constant_p(__B) && __B < 16)
-	lshift = (__v2du) vec_splat_s32(__B);
-      else
-	lshift = (__v2du) vec_splats ((unsigned int) __B);
-
-      result = vec_sl ((__v2di) __A, lshift);
-    }
-
-  return (__m128i) result;
-}
-#endif
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srai_epi16 (__m128i __A, int __B)
-{
-  __v8hu rshift = { 15, 15, 15, 15, 15, 15, 15, 15 };
-  __v8hi result;
-
-  if (__B < 16)
-    {
-      if (__builtin_constant_p(__B))
-	rshift = (__v8hu) vec_splat_s16(__B);
-      else
-	rshift = vec_splats ((unsigned short) __B);
-    }
-  result = vec_sra ((__v8hi) __A, rshift);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srai_epi32 (__m128i __A, int __B)
-{
-  __v4su rshift = { 31, 31, 31, 31 };
-  __v4si result;
-
-  if (__B < 32)
-    {
-      if (__builtin_constant_p(__B))
-	{
-	  if (__B < 16)
-	      rshift = (__v4su) vec_splat_s32(__B);
-	    else
-	      rshift = (__v4su) vec_splats((unsigned int)__B);
-	}
-      else
-	rshift = vec_splats ((unsigned int) __B);
-    }
-  result = vec_sra ((__v4si) __A, rshift);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_bslli_si128 (__m128i __A, const int __N)
-{
-  __v16qu result;
-  const __v16qu zeros = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
-
-  if (__N < 16)
-    result = vec_sld ((__v16qu) __A, zeros, __N);
-  else
-    result = zeros;
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_bsrli_si128 (__m128i __A, const int __N)
-{
-  __v16qu result;
-  const __v16qu zeros = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
-
-  if (__N < 16)
-#ifdef __LITTLE_ENDIAN__
-    if (__builtin_constant_p(__N))
-      /* Would like to use Vector Shift Left Double by Octet
-	 Immediate here to use the immediate form and avoid
-	 load of __N * 8 value into a separate VR.  */
-      result = vec_sld (zeros, (__v16qu) __A, (16 - __N));
-    else
-#endif
-      {
-	__v16qu shift = vec_splats((unsigned char)(__N*8));
-#ifdef __LITTLE_ENDIAN__
-	result = vec_sro ((__v16qu)__A, shift);
-#else
-	result = vec_slo ((__v16qu)__A, shift);
-#endif
-      }
-  else
-    result = zeros;
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srli_si128 (__m128i __A, const int __N)
-{
-  return _mm_bsrli_si128 (__A, __N);
-}
-
-extern __inline  __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_slli_si128 (__m128i __A, const int _imm5)
-{
-  __v16qu result;
-  const __v16qu zeros = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
-
-  if (_imm5 < 16)
-#ifdef __LITTLE_ENDIAN__
-    result = vec_sld ((__v16qu) __A, zeros, _imm5);
-#else
-    result = vec_sld (zeros, (__v16qu) __A, (16 - _imm5));
-#endif
-  else
-    result = zeros;
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-
-_mm_srli_epi16 (__m128i  __A, int __B)
-{
-  __v8hu rshift;
-  __v8hi result = { 0, 0, 0, 0, 0, 0, 0, 0 };
-
-  if (__B < 16)
-    {
-      if (__builtin_constant_p(__B))
-	rshift = (__v8hu) vec_splat_s16(__B);
-      else
-	rshift = vec_splats ((unsigned short) __B);
-
-      result = vec_sr ((__v8hi) __A, rshift);
-    }
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srli_epi32 (__m128i __A, int __B)
-{
-  __v4su rshift;
-  __v4si result = { 0, 0, 0, 0 };
-
-  if (__B < 32)
-    {
-      if (__builtin_constant_p(__B))
-	{
-	  if (__B < 16)
-	      rshift = (__v4su) vec_splat_s32(__B);
-	    else
-	      rshift = (__v4su) vec_splats((unsigned int)__B);
-	}
-      else
-	rshift = vec_splats ((unsigned int) __B);
-
-      result = vec_sr ((__v4si) __A, rshift);
-    }
-
-  return (__m128i) result;
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srli_epi64 (__m128i __A, int __B)
-{
-  __v2du rshift;
-  __v2di result = { 0, 0 };
-
-  if (__B < 64)
-    {
-      if (__builtin_constant_p(__B))
-	{
-	  if (__B < 16)
-	      rshift = (__v2du) vec_splat_s32(__B);
-	    else
-	      rshift = (__v2du) vec_splats((unsigned long long)__B);
-	}
-      else
-	rshift = (__v2du) vec_splats ((unsigned int) __B);
-
-      result = vec_sr ((__v2di) __A, rshift);
-    }
-
-  return (__m128i) result;
-}
-#endif
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sll_epi16 (__m128i __A, __m128i __B)
-{
-  __v8hu lshift;
-  __vector __bool short shmask;
-  const __v8hu shmax = { 15, 15, 15, 15, 15, 15, 15, 15 };
-  __v8hu result;
-
-#ifdef __LITTLE_ENDIAN__
-  lshift = vec_splat ((__v8hu) __B, 0);
-#else
-  lshift = vec_splat ((__v8hu) __B, 3);
-#endif
-  shmask = vec_cmple (lshift, shmax);
-  result = vec_sl ((__v8hu) __A, lshift);
-  result = vec_sel ((__v8hu) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sll_epi32 (__m128i __A, __m128i __B)
-{
-  __v4su lshift;
-  __vector __bool int shmask;
-  const __v4su shmax = { 32, 32, 32, 32 };
-  __v4su result;
-#ifdef __LITTLE_ENDIAN__
-  lshift = vec_splat ((__v4su) __B, 0);
-#else
-  lshift = vec_splat ((__v4su) __B, 1);
-#endif
-  shmask = vec_cmplt (lshift, shmax);
-  result = vec_sl ((__v4su) __A, lshift);
-  result = vec_sel ((__v4su) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sll_epi64 (__m128i __A, __m128i __B)
-{
-  __v2du lshift;
-  __vector __bool long long shmask;
-  const __v2du shmax = { 64, 64 };
-  __v2du result;
-
-  lshift = vec_splat ((__v2du) __B, 0);
-  shmask = vec_cmplt (lshift, shmax);
-  result = vec_sl ((__v2du) __A, lshift);
-  result = vec_sel ((__v2du) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-#endif
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sra_epi16 (__m128i __A, __m128i __B)
-{
-  const __v8hu rshmax = { 15, 15, 15, 15, 15, 15, 15, 15 };
-  __v8hu rshift;
-  __v8hi result;
-
-#ifdef __LITTLE_ENDIAN__
-  rshift = vec_splat ((__v8hu)__B, 0);
-#else
-  rshift = vec_splat ((__v8hu)__B, 3);
-#endif
-  rshift = vec_min (rshift, rshmax);
-  result = vec_sra ((__v8hi) __A, rshift);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sra_epi32 (__m128i __A, __m128i __B)
-{
-  const __v4su rshmax = { 31, 31, 31, 31 };
-  __v4su rshift;
-  __v4si result;
-
-#ifdef __LITTLE_ENDIAN__
-  rshift = vec_splat ((__v4su)__B, 0);
-#else
-  rshift = vec_splat ((__v4su)__B, 1);
-#endif
-  rshift = vec_min (rshift, rshmax);
-  result = vec_sra ((__v4si) __A, rshift);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srl_epi16 (__m128i __A, __m128i __B)
-{
-  __v8hu rshift;
-  __vector __bool short shmask;
-  const __v8hu shmax = { 15, 15, 15, 15, 15, 15, 15, 15 };
-  __v8hu result;
-
-#ifdef __LITTLE_ENDIAN__
-  rshift = vec_splat ((__v8hu) __B, 0);
-#else
-  rshift = vec_splat ((__v8hu) __B, 3);
-#endif
-  shmask = vec_cmple (rshift, shmax);
-  result = vec_sr ((__v8hu) __A, rshift);
-  result = vec_sel ((__v8hu) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srl_epi32 (__m128i __A, __m128i __B)
-{
-  __v4su rshift;
-  __vector __bool int shmask;
-  const __v4su shmax = { 32, 32, 32, 32 };
-  __v4su result;
-
-#ifdef __LITTLE_ENDIAN__
-  rshift = vec_splat ((__v4su) __B, 0);
-#else
-  rshift = vec_splat ((__v4su) __B, 1);
-#endif
-  shmask = vec_cmplt (rshift, shmax);
-  result = vec_sr ((__v4su) __A, rshift);
-  result = vec_sel ((__v4su) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-
-#ifdef _ARCH_PWR8
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_srl_epi64 (__m128i __A, __m128i __B)
-{
-  __v2du rshift;
-  __vector __bool long long shmask;
-  const __v2du shmax = { 64, 64 };
-  __v2du result;
-
-  rshift = vec_splat ((__v2du) __B, 0);
-  shmask = vec_cmplt (rshift, shmax);
-  result = vec_sr ((__v2du) __A, rshift);
-  result = vec_sel ((__v2du) shmask, result, shmask);
-
-  return (__m128i) result;
-}
-#endif
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_and_pd (__m128d __A, __m128d __B)
-{
-  return (vec_and ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_andnot_pd (__m128d __A, __m128d __B)
-{
-  return (vec_andc ((__v2df) __B, (__v2df) __A));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_or_pd (__m128d __A, __m128d __B)
-{
-  return (vec_or ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_xor_pd (__m128d __A, __m128d __B)
-{
-  return (vec_xor ((__v2df) __A, (__v2df) __B));
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_and_si128 (__m128i __A, __m128i __B)
-{
-  return (__m128i)vec_and ((__v2di) __A, (__v2di) __B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_andnot_si128 (__m128i __A, __m128i __B)
-{
-  return (__m128i)vec_andc ((__v2di) __B, (__v2di) __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_or_si128 (__m128i __A, __m128i __B)
-{
-  return (__m128i)vec_or ((__v2di) __A, (__v2di) __B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_xor_si128 (__m128i __A, __m128i __B)
-{
-  return (__m128i)vec_xor ((__v2di) __A, (__v2di) __B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpeq ((__v16qi) __A, (__v16qi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpeq ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpeq ((__v4si) __A, (__v4si)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmplt ((__v16qi) __A, (__v16qi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmplt ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmplt ((__v4si) __A, (__v4si)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_epi8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpgt ((__v16qi) __A, (__v16qi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpgt ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_epi32 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_cmpgt ((__v4si) __A, (__v4si)__B);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_extract_epi16 (__m128i const __A, int const __N)
-{
-  return (unsigned short) ((__v8hi)__A)[__N & 7];
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_insert_epi16 (__m128i const __A, int const __D, int const __N)
-{
-  __v8hi result = (__v8hi)__A;
-
-  result [(__N & 7)] = __D;
-
-  return (__m128i) result;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_max ((__v8hi)__A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_epu8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_max ((__v16qu) __A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_epi16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_min ((__v8hi) __A, (__v8hi)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_epu8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_min ((__v16qu) __A, (__v16qu)__B);
-}
-
-
-#ifdef _ARCH_PWR8
-/* Intrinsic functions that require PowerISA 2.07 minimum.  */
-
-/* Creates a 4-bit mask from the most significant bits of the SPFP values.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movemask_epi8 (__m128i __A)
-{
-  __vector unsigned long long result;
-  static const __vector unsigned char perm_mask =
-    {
-	0x78, 0x70, 0x68, 0x60, 0x58, 0x50, 0x48, 0x40,
-	0x38, 0x30, 0x28, 0x20, 0x18, 0x10, 0x08, 0x00
-    };
-
-  result = ((__vector unsigned long long)
-	    vec_vbpermq ((__vector unsigned char) __A,
-			 (__vector unsigned char) perm_mask));
-
-#ifdef __LITTLE_ENDIAN__
-  return result[1];
-#else
-  return result[0];
-#endif
-}
-#endif /* _ARCH_PWR8 */
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mulhi_epu16 (__m128i __A, __m128i __B)
-{
-  __v4su w0, w1;
-  __v16qu xform1 = {
-#ifdef __LITTLE_ENDIAN__
-      0x02, 0x03, 0x12, 0x13,  0x06, 0x07, 0x16, 0x17,
-      0x0A, 0x0B, 0x1A, 0x1B,  0x0E, 0x0F, 0x1E, 0x1F
-#else
-      0x00, 0x01, 0x10, 0x11,  0x04, 0x05, 0x14, 0x15,
-      0x08, 0x09, 0x18, 0x19,  0x0C, 0x0D, 0x1C, 0x1D
-#endif
-    };
-
-  w0 = vec_vmuleuh ((__v8hu)__A, (__v8hu)__B);
-  w1 = vec_vmulouh ((__v8hu)__A, (__v8hu)__B);
-  return (__m128i) vec_perm (w0, w1, xform1);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_shufflehi_epi16 (__m128i __A, const int __mask)
-{
-  unsigned long element_selector_98 = __mask & 0x03;
-  unsigned long element_selector_BA = (__mask >> 2) & 0x03;
-  unsigned long element_selector_DC = (__mask >> 4) & 0x03;
-  unsigned long element_selector_FE = (__mask >> 6) & 0x03;
-  static const unsigned short permute_selectors[4] =
-    {
-#ifdef __LITTLE_ENDIAN__
-	      0x0908, 0x0B0A, 0x0D0C, 0x0F0E
-#else
-	      0x0809, 0x0A0B, 0x0C0D, 0x0E0F
-#endif
-    };
-  __v2du pmask =
-#ifdef __LITTLE_ENDIAN__
-      { 0x1716151413121110UL,  0UL};
-#else
-      { 0x1011121314151617UL,  0UL};
-#endif
-  __m64_union t;
-  __v2du a, r;
-
-  t.as_short[0] = permute_selectors[element_selector_98];
-  t.as_short[1] = permute_selectors[element_selector_BA];
-  t.as_short[2] = permute_selectors[element_selector_DC];
-  t.as_short[3] = permute_selectors[element_selector_FE];
-  pmask[1] = t.as_m64;
-  a = (__v2du)__A;
-  r = vec_perm (a, a, (__vector unsigned char)pmask);
-  return (__m128i) r;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_shufflelo_epi16 (__m128i __A, const int __mask)
-{
-  unsigned long element_selector_10 = __mask & 0x03;
-  unsigned long element_selector_32 = (__mask >> 2) & 0x03;
-  unsigned long element_selector_54 = (__mask >> 4) & 0x03;
-  unsigned long element_selector_76 = (__mask >> 6) & 0x03;
-  static const unsigned short permute_selectors[4] =
-    {
-#ifdef __LITTLE_ENDIAN__
-	      0x0100, 0x0302, 0x0504, 0x0706
-#else
-	      0x0001, 0x0203, 0x0405, 0x0607
-#endif
-    };
-  __v2du pmask =
-#ifdef __LITTLE_ENDIAN__
-                 { 0UL,  0x1f1e1d1c1b1a1918UL};
-#else
-                 { 0UL,  0x18191a1b1c1d1e1fUL};
-#endif
-  __m64_union t;
-  __v2du a, r;
-  t.as_short[0] = permute_selectors[element_selector_10];
-  t.as_short[1] = permute_selectors[element_selector_32];
-  t.as_short[2] = permute_selectors[element_selector_54];
-  t.as_short[3] = permute_selectors[element_selector_76];
-  pmask[0] = t.as_m64;
-  a = (__v2du)__A;
-  r = vec_perm (a, a, (__vector unsigned char)pmask);
-  return (__m128i) r;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_shuffle_epi32 (__m128i __A, const int __mask)
-{
-  unsigned long element_selector_10 = __mask & 0x03;
-  unsigned long element_selector_32 = (__mask >> 2) & 0x03;
-  unsigned long element_selector_54 = (__mask >> 4) & 0x03;
-  unsigned long element_selector_76 = (__mask >> 6) & 0x03;
-  static const unsigned int permute_selectors[4] =
-    {
-#ifdef __LITTLE_ENDIAN__
-	0x03020100, 0x07060504, 0x0B0A0908, 0x0F0E0D0C
-#else
-      0x00010203, 0x04050607, 0x08090A0B, 0x0C0D0E0F
-#endif
-    };
-  __v4su t;
-
-  t[0] = permute_selectors[element_selector_10];
-  t[1] = permute_selectors[element_selector_32];
-  t[2] = permute_selectors[element_selector_54] + 0x10101010;
-  t[3] = permute_selectors[element_selector_76] + 0x10101010;
-  return (__m128i)vec_perm ((__v4si) __A, (__v4si)__A, (__vector unsigned char)t);
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_maskmoveu_si128 (__m128i __A, __m128i __B, char *__C)
-{
-  __v2du hibit = { 0x7f7f7f7f7f7f7f7fUL, 0x7f7f7f7f7f7f7f7fUL};
-  __v16qu mask, tmp;
-  __m128i_u *p = (__m128i_u*)__C;
-
-  tmp = (__v16qu)_mm_loadu_si128(p);
-  mask = (__v16qu)vec_cmpgt ((__v16qu)__B, (__v16qu)hibit);
-  tmp = vec_sel (tmp, (__v16qu)__A, mask);
-  _mm_storeu_si128 (p, (__m128i)tmp);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_avg_epu8 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_avg ((__v16qu)__A, (__v16qu)__B);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_avg_epu16 (__m128i __A, __m128i __B)
-{
-  return (__m128i) vec_avg ((__v8hu)__A, (__v8hu)__B);
-}
-
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sad_epu8 (__m128i __A, __m128i __B)
-{
-  __v16qu a, b;
-  __v16qu vmin, vmax, vabsdiff;
-  __v4si vsum;
-  const __v4su zero = { 0, 0, 0, 0 };
-  __v4si result;
-
-  a = (__v16qu) __A;
-  b = (__v16qu) __B;
-  vmin = vec_min (a, b);
-  vmax = vec_max (a, b);
-  vabsdiff = vec_sub (vmax, vmin);
-  /* Sum four groups of bytes into integers.  */
-  vsum = (__vector signed int) vec_sum4s (vabsdiff, zero);
-  /* Sum across four integers with two integer results.  */
-  result = vec_sum2s (vsum, (__vector signed int) zero);
-  /* Rotate the sums into the correct position.  */
-#ifdef __LITTLE_ENDIAN__
-  result = vec_sld (result, result, 4);
-#else
-  result = vec_sld (result, result, 6);
-#endif
-  /* Rotate the sums into the correct position.  */
-  return (__m128i) result;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_si32 (int *__A, int __B)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "dcbtstt 0,%0"
-    :
-    : "b" (__A)
-    : "memory"
-  );
-  *__A = __B;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_si64 (long long int *__A, long long int __B)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "	dcbtstt	0,%0"
-    :
-    : "b" (__A)
-    : "memory"
-  );
-  *__A = __B;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_si128 (__m128i *__A, __m128i __B)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "dcbtstt 0,%0"
-    :
-    : "b" (__A)
-    : "memory"
-  );
-  *__A = __B;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_pd (double *__A, __m128d __B)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "dcbtstt 0,%0"
-    :
-    : "b" (__A)
-    : "memory"
-  );
-  *(__m128d*)__A = __B;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_clflush (void const *__A)
-{
-  /* Use the data cache block flush.  */
-  __asm__ (
-    "dcbf 0,%0"
-    :
-    : "b" (__A)
-    : "memory"
-  );
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_lfence (void)
-{
-  /* Use light weight sync for load to load ordering.  */
-  __atomic_thread_fence (__ATOMIC_RELEASE);
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mfence (void)
-{
-  /* Use heavy weight sync for any to any ordering.  */
-  __atomic_thread_fence (__ATOMIC_SEQ_CST);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi32_si128 (int __A)
-{
-  return _mm_set_epi32 (0, 0, 0, __A);
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64_si128 (long long __A)
-{
-  return __extension__ (__m128i)(__v2di){ __A, 0LL };
-}
-
-/* Microsoft intrinsic.  */
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64x_si128 (long long __A)
-{
-  return __extension__ (__m128i)(__v2di){ __A, 0LL };
-}
-
-/* Casts between various SP, DP, INT vector types.  Note that these do no
-   conversion of values, they just change the type.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castpd_ps(__m128d __A)
-{
-  return (__m128) __A;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castpd_si128(__m128d __A)
-{
-  return (__m128i) __A;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castps_pd(__m128 __A)
-{
-  return (__m128d) __A;
-}
-
-extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castps_si128(__m128 __A)
-{
-  return (__m128i) __A;
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castsi128_ps(__m128i __A)
-{
-  return (__m128) __A;
-}
-
-extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_castsi128_pd(__m128i __A)
-{
-  return (__m128d) __A;
-}
-
-#endif /* EMMINTRIN_H_ */
diff --git a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h b/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
deleted file mode 100644
index 3658919..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mm_malloc.h
+++ /dev/null
@@ -1,48 +0,0 @@
-/*===---- mm_malloc.h - Implementation of _mm_malloc and _mm_free ----------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef _MM_MALLOC_H_INCLUDED
-#define _MM_MALLOC_H_INCLUDED
-
-#include <stdlib.h>
-
-/* We can't depend on <stdlib.h> since the prototype of posix_memalign
-   may not be visible.  */
-#ifndef __cplusplus
-extern int posix_memalign (void **, size_t, size_t);
-#else
-extern "C" int posix_memalign (void **, size_t, size_t) throw ();
-#endif
-
-static __inline void *
-_mm_malloc (size_t size, size_t alignment)
-{
-  /* PowerPC64 ELF V2 ABI requires quadword alignment.  */
-  size_t vec_align = sizeof (__vector float);
-  /* Linux GLIBC malloc alignment is at least 2 X ptr size.  */
-  size_t malloc_align = (sizeof (void *) + sizeof (void *));
-  void *ptr;
-
-  if (alignment == malloc_align && alignment == vec_align)
-    return malloc (size);
-  if (alignment < vec_align)
-    alignment = vec_align;
-  if (posix_memalign (&ptr, alignment, size) == 0)
-    return ptr;
-  else
-    return NULL;
-}
-
-static __inline void
-_mm_free (void * ptr)
-{
-  free (ptr);
-}
-
-#endif /* _MM_MALLOC_H_INCLUDED */
diff --git a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h b/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
deleted file mode 100644
index b949653..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/mmintrin.h
+++ /dev/null
@@ -1,1443 +0,0 @@
-/*===---- mmintrin.h - Implementation of MMX intrinsics on PowerPC ---------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-/* Implemented from the specification included in the Intel C++ Compiler
-   User Guide and Reference, version 9.0.  */
-
-#ifndef NO_WARN_X86_INTRINSICS
-/* This header file is to help porting code using Intel intrinsics
-   explicitly from x86_64 to powerpc64/powerpc64le.
-
-   Since PowerPC target doesn't support native 64-bit vector type, we
-   typedef __m64 to 64-bit unsigned long long in MMX intrinsics, which
-   works well for _si64 and some _pi32 operations.
-
-   For _pi16 and _pi8 operations, it's better to transfer __m64 into
-   128-bit PowerPC vector first. Power8 introduced direct register
-   move instructions which helps for more efficient implementation.
-
-   It's user's responsibility to determine if the results of such port
-   are acceptable or further changes are needed. Please note that much
-   code using Intel intrinsics CAN BE REWRITTEN in more portable and
-   efficient standard C or GNU C extensions with 64-bit scalar
-   operations, or 128-bit SSE/Altivec operations, which are more
-   recommended. */
-#error                                                                         \
-    "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this error."
-#endif
-
-#ifndef _MMINTRIN_H_INCLUDED
-#define _MMINTRIN_H_INCLUDED
-
-#include <altivec.h>
-/* The Intel API is flexible enough that we must allow aliasing with other
-   vector types, and their scalar components.  */
-typedef __attribute__((__aligned__(8))) unsigned long long __m64;
-
-typedef __attribute__((__aligned__(8))) union {
-  __m64 as_m64;
-  char as_char[8];
-  signed char as_signed_char[8];
-  short as_short[4];
-  int as_int[2];
-  long long as_long_long;
-  float as_float[2];
-  double as_double;
-} __m64_union;
-
-/* Empty the multimedia state.  */
-extern __inline void
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_empty(void) {
-  /* nothing to do on PowerPC.  */
-}
-
-extern __inline void
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_empty(void) {
-  /* nothing to do on PowerPC.  */
-}
-
-/* Convert I to a __m64 object.  The integer is zero-extended to 64-bits.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtsi32_si64(int __i) {
-  return (__m64)(unsigned int)__i;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_from_int(int __i) {
-  return _mm_cvtsi32_si64(__i);
-}
-
-/* Convert the lower 32 bits of the __m64 object into an integer.  */
-extern __inline int
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtsi64_si32(__m64 __i) {
-  return ((int)__i);
-}
-
-extern __inline int
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_to_int(__m64 __i) {
-  return _mm_cvtsi64_si32(__i);
-}
-
-/* Convert I to a __m64 object.  */
-
-/* Intel intrinsic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_from_int64(long long __i) {
-  return (__m64)__i;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtsi64_m64(long long __i) {
-  return (__m64)__i;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtsi64x_si64(long long __i) {
-  return (__m64)__i;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set_pi64x(long long __i) {
-  return (__m64)__i;
-}
-
-/* Convert the __m64 object to a 64bit integer.  */
-
-/* Intel intrinsic.  */
-extern __inline long long
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_to_int64(__m64 __i) {
-  return (long long)__i;
-}
-
-extern __inline long long
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtm64_si64(__m64 __i) {
-  return (long long)__i;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cvtsi64_si64x(__m64 __i) {
-  return (long long)__i;
-}
-
-#ifdef _ARCH_PWR8
-/* Pack the four 16-bit values from M1 into the lower four 8-bit values of
-   the result, and the four 16-bit values from M2 into the upper four 8-bit
-   values of the result, all with signed saturation.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_packs_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short vm1;
-  __vector signed char vresult;
-
-  vm1 = (__vector signed short)(__vector unsigned long long)
-#ifdef __LITTLE_ENDIAN__
-      {__m1, __m2};
-#else
-      {__m2, __m1};
-#endif
-  vresult = vec_packs(vm1, vm1);
-  return (__m64)((__vector long long)vresult)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_packsswb(__m64 __m1, __m64 __m2) {
-  return _mm_packs_pi16(__m1, __m2);
-}
-
-/* Pack the two 32-bit values from M1 in to the lower two 16-bit values of
-   the result, and the two 32-bit values from M2 into the upper two 16-bit
-   values of the result, all with signed saturation.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_packs_pi32(__m64 __m1, __m64 __m2) {
-  __vector signed int vm1;
-  __vector signed short vresult;
-
-  vm1 = (__vector signed int)(__vector unsigned long long)
-#ifdef __LITTLE_ENDIAN__
-      {__m1, __m2};
-#else
-      {__m2, __m1};
-#endif
-  vresult = vec_packs(vm1, vm1);
-  return (__m64)((__vector long long)vresult)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_packssdw(__m64 __m1, __m64 __m2) {
-  return _mm_packs_pi32(__m1, __m2);
-}
-
-/* Pack the four 16-bit values from M1 into the lower four 8-bit values of
-   the result, and the four 16-bit values from M2 into the upper four 8-bit
-   values of the result, all with unsigned saturation.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_packs_pu16(__m64 __m1, __m64 __m2) {
-  __vector unsigned char r;
-  __vector signed short vm1 = (__vector signed short)(__vector long long)
-#ifdef __LITTLE_ENDIAN__
-      {__m1, __m2};
-#else
-      {__m2, __m1};
-#endif
-  const __vector signed short __zero = {0};
-  __vector __bool short __select = vec_cmplt(vm1, __zero);
-  r = vec_packs((__vector unsigned short)vm1, (__vector unsigned short)vm1);
-  __vector __bool char packsel = vec_pack(__select, __select);
-  r = vec_sel(r, (const __vector unsigned char)__zero, packsel);
-  return (__m64)((__vector long long)r)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_packuswb(__m64 __m1, __m64 __m2) {
-  return _mm_packs_pu16(__m1, __m2);
-}
-#endif /* end ARCH_PWR8 */
-
-/* Interleave the four 8-bit values from the high half of M1 with the four
-   8-bit values from the high half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpackhi_pi8(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector unsigned char a, b, c;
-
-  a = (__vector unsigned char)vec_splats(__m1);
-  b = (__vector unsigned char)vec_splats(__m2);
-  c = vec_mergel(a, b);
-  return (__m64)((__vector long long)c)[1];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = m1.as_char[4];
-  res.as_char[1] = m2.as_char[4];
-  res.as_char[2] = m1.as_char[5];
-  res.as_char[3] = m2.as_char[5];
-  res.as_char[4] = m1.as_char[6];
-  res.as_char[5] = m2.as_char[6];
-  res.as_char[6] = m1.as_char[7];
-  res.as_char[7] = m2.as_char[7];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpckhbw(__m64 __m1, __m64 __m2) {
-  return _mm_unpackhi_pi8(__m1, __m2);
-}
-
-/* Interleave the two 16-bit values from the high half of M1 with the two
-   16-bit values from the high half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpackhi_pi16(__m64 __m1, __m64 __m2) {
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = m1.as_short[2];
-  res.as_short[1] = m2.as_short[2];
-  res.as_short[2] = m1.as_short[3];
-  res.as_short[3] = m2.as_short[3];
-
-  return (__m64)res.as_m64;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpckhwd(__m64 __m1, __m64 __m2) {
-  return _mm_unpackhi_pi16(__m1, __m2);
-}
-/* Interleave the 32-bit value from the high half of M1 with the 32-bit
-   value from the high half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpackhi_pi32(__m64 __m1, __m64 __m2) {
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = m1.as_int[1];
-  res.as_int[1] = m2.as_int[1];
-
-  return (__m64)res.as_m64;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpckhdq(__m64 __m1, __m64 __m2) {
-  return _mm_unpackhi_pi32(__m1, __m2);
-}
-/* Interleave the four 8-bit values from the low half of M1 with the four
-   8-bit values from the low half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpacklo_pi8(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector unsigned char a, b, c;
-
-  a = (__vector unsigned char)vec_splats(__m1);
-  b = (__vector unsigned char)vec_splats(__m2);
-  c = vec_mergel(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = m1.as_char[0];
-  res.as_char[1] = m2.as_char[0];
-  res.as_char[2] = m1.as_char[1];
-  res.as_char[3] = m2.as_char[1];
-  res.as_char[4] = m1.as_char[2];
-  res.as_char[5] = m2.as_char[2];
-  res.as_char[6] = m1.as_char[3];
-  res.as_char[7] = m2.as_char[3];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpcklbw(__m64 __m1, __m64 __m2) {
-  return _mm_unpacklo_pi8(__m1, __m2);
-}
-/* Interleave the two 16-bit values from the low half of M1 with the two
-   16-bit values from the low half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpacklo_pi16(__m64 __m1, __m64 __m2) {
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = m1.as_short[0];
-  res.as_short[1] = m2.as_short[0];
-  res.as_short[2] = m1.as_short[1];
-  res.as_short[3] = m2.as_short[1];
-
-  return (__m64)res.as_m64;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpcklwd(__m64 __m1, __m64 __m2) {
-  return _mm_unpacklo_pi16(__m1, __m2);
-}
-
-/* Interleave the 32-bit value from the low half of M1 with the 32-bit
-   value from the low half of M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_unpacklo_pi32(__m64 __m1, __m64 __m2) {
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = m1.as_int[0];
-  res.as_int[1] = m2.as_int[0];
-
-  return (__m64)res.as_m64;
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_punpckldq(__m64 __m1, __m64 __m2) {
-  return _mm_unpacklo_pi32(__m1, __m2);
-}
-
-/* Add the 8-bit values in M1 to the 8-bit values in M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_add_pi8(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed char a, b, c;
-
-  a = (__vector signed char)vec_splats(__m1);
-  b = (__vector signed char)vec_splats(__m2);
-  c = vec_add(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = m1.as_char[0] + m2.as_char[0];
-  res.as_char[1] = m1.as_char[1] + m2.as_char[1];
-  res.as_char[2] = m1.as_char[2] + m2.as_char[2];
-  res.as_char[3] = m1.as_char[3] + m2.as_char[3];
-  res.as_char[4] = m1.as_char[4] + m2.as_char[4];
-  res.as_char[5] = m1.as_char[5] + m2.as_char[5];
-  res.as_char[6] = m1.as_char[6] + m2.as_char[6];
-  res.as_char[7] = m1.as_char[7] + m2.as_char[7];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddb(__m64 __m1, __m64 __m2) {
-  return _mm_add_pi8(__m1, __m2);
-}
-
-/* Add the 16-bit values in M1 to the 16-bit values in M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_add_pi16(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = vec_add(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = m1.as_short[0] + m2.as_short[0];
-  res.as_short[1] = m1.as_short[1] + m2.as_short[1];
-  res.as_short[2] = m1.as_short[2] + m2.as_short[2];
-  res.as_short[3] = m1.as_short[3] + m2.as_short[3];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddw(__m64 __m1, __m64 __m2) {
-  return _mm_add_pi16(__m1, __m2);
-}
-
-/* Add the 32-bit values in M1 to the 32-bit values in M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_add_pi32(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR9
-  __vector signed int a, b, c;
-
-  a = (__vector signed int)vec_splats(__m1);
-  b = (__vector signed int)vec_splats(__m2);
-  c = vec_add(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = m1.as_int[0] + m2.as_int[0];
-  res.as_int[1] = m1.as_int[1] + m2.as_int[1];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddd(__m64 __m1, __m64 __m2) {
-  return _mm_add_pi32(__m1, __m2);
-}
-
-/* Subtract the 8-bit values in M2 from the 8-bit values in M1.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sub_pi8(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed char a, b, c;
-
-  a = (__vector signed char)vec_splats(__m1);
-  b = (__vector signed char)vec_splats(__m2);
-  c = vec_sub(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = m1.as_char[0] - m2.as_char[0];
-  res.as_char[1] = m1.as_char[1] - m2.as_char[1];
-  res.as_char[2] = m1.as_char[2] - m2.as_char[2];
-  res.as_char[3] = m1.as_char[3] - m2.as_char[3];
-  res.as_char[4] = m1.as_char[4] - m2.as_char[4];
-  res.as_char[5] = m1.as_char[5] - m2.as_char[5];
-  res.as_char[6] = m1.as_char[6] - m2.as_char[6];
-  res.as_char[7] = m1.as_char[7] - m2.as_char[7];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubb(__m64 __m1, __m64 __m2) {
-  return _mm_sub_pi8(__m1, __m2);
-}
-
-/* Subtract the 16-bit values in M2 from the 16-bit values in M1.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sub_pi16(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = vec_sub(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = m1.as_short[0] - m2.as_short[0];
-  res.as_short[1] = m1.as_short[1] - m2.as_short[1];
-  res.as_short[2] = m1.as_short[2] - m2.as_short[2];
-  res.as_short[3] = m1.as_short[3] - m2.as_short[3];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubw(__m64 __m1, __m64 __m2) {
-  return _mm_sub_pi16(__m1, __m2);
-}
-
-/* Subtract the 32-bit values in M2 from the 32-bit values in M1.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sub_pi32(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR9
-  __vector signed int a, b, c;
-
-  a = (__vector signed int)vec_splats(__m1);
-  b = (__vector signed int)vec_splats(__m2);
-  c = vec_sub(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = m1.as_int[0] - m2.as_int[0];
-  res.as_int[1] = m1.as_int[1] - m2.as_int[1];
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubd(__m64 __m1, __m64 __m2) {
-  return _mm_sub_pi32(__m1, __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_add_si64(__m64 __m1, __m64 __m2) {
-  return (__m1 + __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sub_si64(__m64 __m1, __m64 __m2) {
-  return (__m1 - __m2);
-}
-
-/* Shift the 64-bit value in M left by COUNT.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sll_si64(__m64 __m, __m64 __count) {
-  return (__m << __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psllq(__m64 __m, __m64 __count) {
-  return _mm_sll_si64(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_slli_si64(__m64 __m, const int __count) {
-  return (__m << __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psllqi(__m64 __m, const int __count) {
-  return _mm_slli_si64(__m, __count);
-}
-
-/* Shift the 64-bit value in M left by COUNT; shift in zeros.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srl_si64(__m64 __m, __m64 __count) {
-  return (__m >> __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrlq(__m64 __m, __m64 __count) {
-  return _mm_srl_si64(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srli_si64(__m64 __m, const int __count) {
-  return (__m >> __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrlqi(__m64 __m, const int __count) {
-  return _mm_srli_si64(__m, __count);
-}
-
-/* Bit-wise AND the 64-bit values in M1 and M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_and_si64(__m64 __m1, __m64 __m2) {
-  return (__m1 & __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pand(__m64 __m1, __m64 __m2) {
-  return _mm_and_si64(__m1, __m2);
-}
-
-/* Bit-wise complement the 64-bit value in M1 and bit-wise AND it with the
-   64-bit value in M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_andnot_si64(__m64 __m1, __m64 __m2) {
-  return (~__m1 & __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pandn(__m64 __m1, __m64 __m2) {
-  return _mm_andnot_si64(__m1, __m2);
-}
-
-/* Bit-wise inclusive OR the 64-bit values in M1 and M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_or_si64(__m64 __m1, __m64 __m2) {
-  return (__m1 | __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_por(__m64 __m1, __m64 __m2) {
-  return _mm_or_si64(__m1, __m2);
-}
-
-/* Bit-wise exclusive OR the 64-bit values in M1 and M2.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_xor_si64(__m64 __m1, __m64 __m2) {
-  return (__m1 ^ __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pxor(__m64 __m1, __m64 __m2) {
-  return _mm_xor_si64(__m1, __m2);
-}
-
-/* Creates a 64-bit zero.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_setzero_si64(void) {
-  return (__m64)0;
-}
-
-/* Compare eight 8-bit values.  The result of the comparison is 0xFF if the
-   test is true and zero if false.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpeq_pi8(__m64 __m1, __m64 __m2) {
-#if defined(_ARCH_PWR6) && defined(__powerpc64__)
-  __m64 res;
-  __asm__("cmpb %0,%1,%2;\n" : "=r"(res) : "r"(__m1), "r"(__m2) :);
-  return (res);
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = (m1.as_char[0] == m2.as_char[0]) ? -1 : 0;
-  res.as_char[1] = (m1.as_char[1] == m2.as_char[1]) ? -1 : 0;
-  res.as_char[2] = (m1.as_char[2] == m2.as_char[2]) ? -1 : 0;
-  res.as_char[3] = (m1.as_char[3] == m2.as_char[3]) ? -1 : 0;
-  res.as_char[4] = (m1.as_char[4] == m2.as_char[4]) ? -1 : 0;
-  res.as_char[5] = (m1.as_char[5] == m2.as_char[5]) ? -1 : 0;
-  res.as_char[6] = (m1.as_char[6] == m2.as_char[6]) ? -1 : 0;
-  res.as_char[7] = (m1.as_char[7] == m2.as_char[7]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpeqb(__m64 __m1, __m64 __m2) {
-  return _mm_cmpeq_pi8(__m1, __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpgt_pi8(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed char a, b, c;
-
-  a = (__vector signed char)vec_splats(__m1);
-  b = (__vector signed char)vec_splats(__m2);
-  c = (__vector signed char)vec_cmpgt(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_char[0] = (m1.as_char[0] > m2.as_char[0]) ? -1 : 0;
-  res.as_char[1] = (m1.as_char[1] > m2.as_char[1]) ? -1 : 0;
-  res.as_char[2] = (m1.as_char[2] > m2.as_char[2]) ? -1 : 0;
-  res.as_char[3] = (m1.as_char[3] > m2.as_char[3]) ? -1 : 0;
-  res.as_char[4] = (m1.as_char[4] > m2.as_char[4]) ? -1 : 0;
-  res.as_char[5] = (m1.as_char[5] > m2.as_char[5]) ? -1 : 0;
-  res.as_char[6] = (m1.as_char[6] > m2.as_char[6]) ? -1 : 0;
-  res.as_char[7] = (m1.as_char[7] > m2.as_char[7]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpgtb(__m64 __m1, __m64 __m2) {
-  return _mm_cmpgt_pi8(__m1, __m2);
-}
-
-/* Compare four 16-bit values.  The result of the comparison is 0xFFFF if
-   the test is true and zero if false.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpeq_pi16(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = (__vector signed short)vec_cmpeq(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = (m1.as_short[0] == m2.as_short[0]) ? -1 : 0;
-  res.as_short[1] = (m1.as_short[1] == m2.as_short[1]) ? -1 : 0;
-  res.as_short[2] = (m1.as_short[2] == m2.as_short[2]) ? -1 : 0;
-  res.as_short[3] = (m1.as_short[3] == m2.as_short[3]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpeqw(__m64 __m1, __m64 __m2) {
-  return _mm_cmpeq_pi16(__m1, __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpgt_pi16(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR8
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = (__vector signed short)vec_cmpgt(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_short[0] = (m1.as_short[0] > m2.as_short[0]) ? -1 : 0;
-  res.as_short[1] = (m1.as_short[1] > m2.as_short[1]) ? -1 : 0;
-  res.as_short[2] = (m1.as_short[2] > m2.as_short[2]) ? -1 : 0;
-  res.as_short[3] = (m1.as_short[3] > m2.as_short[3]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpgtw(__m64 __m1, __m64 __m2) {
-  return _mm_cmpgt_pi16(__m1, __m2);
-}
-
-/* Compare two 32-bit values.  The result of the comparison is 0xFFFFFFFF if
-   the test is true and zero if false.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpeq_pi32(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR9
-  __vector signed int a, b, c;
-
-  a = (__vector signed int)vec_splats(__m1);
-  b = (__vector signed int)vec_splats(__m2);
-  c = (__vector signed int)vec_cmpeq(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = (m1.as_int[0] == m2.as_int[0]) ? -1 : 0;
-  res.as_int[1] = (m1.as_int[1] == m2.as_int[1]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpeqd(__m64 __m1, __m64 __m2) {
-  return _mm_cmpeq_pi32(__m1, __m2);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_cmpgt_pi32(__m64 __m1, __m64 __m2) {
-#if _ARCH_PWR9
-  __vector signed int a, b, c;
-
-  a = (__vector signed int)vec_splats(__m1);
-  b = (__vector signed int)vec_splats(__m2);
-  c = (__vector signed int)vec_cmpgt(a, b);
-  return (__m64)((__vector long long)c)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __m1;
-  m2.as_m64 = __m2;
-
-  res.as_int[0] = (m1.as_int[0] > m2.as_int[0]) ? -1 : 0;
-  res.as_int[1] = (m1.as_int[1] > m2.as_int[1]) ? -1 : 0;
-
-  return (__m64)res.as_m64;
-#endif
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pcmpgtd(__m64 __m1, __m64 __m2) {
-  return _mm_cmpgt_pi32(__m1, __m2);
-}
-
-#if _ARCH_PWR8
-/* Add the 8-bit values in M1 to the 8-bit values in M2 using signed
-   saturated arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_adds_pi8(__m64 __m1, __m64 __m2) {
-  __vector signed char a, b, c;
-
-  a = (__vector signed char)vec_splats(__m1);
-  b = (__vector signed char)vec_splats(__m2);
-  c = vec_adds(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddsb(__m64 __m1, __m64 __m2) {
-  return _mm_adds_pi8(__m1, __m2);
-}
-/* Add the 16-bit values in M1 to the 16-bit values in M2 using signed
-   saturated arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_adds_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = vec_adds(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddsw(__m64 __m1, __m64 __m2) {
-  return _mm_adds_pi16(__m1, __m2);
-}
-/* Add the 8-bit values in M1 to the 8-bit values in M2 using unsigned
-   saturated arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_adds_pu8(__m64 __m1, __m64 __m2) {
-  __vector unsigned char a, b, c;
-
-  a = (__vector unsigned char)vec_splats(__m1);
-  b = (__vector unsigned char)vec_splats(__m2);
-  c = vec_adds(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddusb(__m64 __m1, __m64 __m2) {
-  return _mm_adds_pu8(__m1, __m2);
-}
-
-/* Add the 16-bit values in M1 to the 16-bit values in M2 using unsigned
-   saturated arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_adds_pu16(__m64 __m1, __m64 __m2) {
-  __vector unsigned short a, b, c;
-
-  a = (__vector unsigned short)vec_splats(__m1);
-  b = (__vector unsigned short)vec_splats(__m2);
-  c = vec_adds(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_paddusw(__m64 __m1, __m64 __m2) {
-  return _mm_adds_pu16(__m1, __m2);
-}
-
-/* Subtract the 8-bit values in M2 from the 8-bit values in M1 using signed
-   saturating arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_subs_pi8(__m64 __m1, __m64 __m2) {
-  __vector signed char a, b, c;
-
-  a = (__vector signed char)vec_splats(__m1);
-  b = (__vector signed char)vec_splats(__m2);
-  c = vec_subs(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubsb(__m64 __m1, __m64 __m2) {
-  return _mm_subs_pi8(__m1, __m2);
-}
-
-/* Subtract the 16-bit values in M2 from the 16-bit values in M1 using
-   signed saturating arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_subs_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = vec_subs(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubsw(__m64 __m1, __m64 __m2) {
-  return _mm_subs_pi16(__m1, __m2);
-}
-
-/* Subtract the 8-bit values in M2 from the 8-bit values in M1 using
-   unsigned saturating arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_subs_pu8(__m64 __m1, __m64 __m2) {
-  __vector unsigned char a, b, c;
-
-  a = (__vector unsigned char)vec_splats(__m1);
-  b = (__vector unsigned char)vec_splats(__m2);
-  c = vec_subs(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubusb(__m64 __m1, __m64 __m2) {
-  return _mm_subs_pu8(__m1, __m2);
-}
-
-/* Subtract the 16-bit values in M2 from the 16-bit values in M1 using
-   unsigned saturating arithmetic.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_subs_pu16(__m64 __m1, __m64 __m2) {
-  __vector unsigned short a, b, c;
-
-  a = (__vector unsigned short)vec_splats(__m1);
-  b = (__vector unsigned short)vec_splats(__m2);
-  c = vec_subs(a, b);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psubusw(__m64 __m1, __m64 __m2) {
-  return _mm_subs_pu16(__m1, __m2);
-}
-
-/* Multiply four 16-bit values in M1 by four 16-bit values in M2 producing
-   four 32-bit intermediate results, which are then summed by pairs to
-   produce two 32-bit results.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_madd_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short a, b;
-  __vector signed int c;
-  __vector signed int zero = {0, 0, 0, 0};
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = vec_vmsumshm(a, b, zero);
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pmaddwd(__m64 __m1, __m64 __m2) {
-  return _mm_madd_pi16(__m1, __m2);
-}
-/* Multiply four signed 16-bit values in M1 by four signed 16-bit values in
-   M2 and produce the high 16 bits of the 32-bit results.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_mulhi_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short a, b;
-  __vector signed short c;
-  __vector signed int w0, w1;
-  __vector unsigned char xform1 = {
-#ifdef __LITTLE_ENDIAN__
-      0x02, 0x03, 0x12, 0x13, 0x06, 0x07, 0x16, 0x17, 0x0A,
-      0x0B, 0x1A, 0x1B, 0x0E, 0x0F, 0x1E, 0x1F
-#else
-      0x00, 0x01, 0x10, 0x11, 0x04, 0x05, 0x14, 0x15, 0x00,
-      0x01, 0x10, 0x11, 0x04, 0x05, 0x14, 0x15
-#endif
-  };
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-
-  w0 = vec_vmulesh(a, b);
-  w1 = vec_vmulosh(a, b);
-  c = (__vector signed short)vec_perm(w0, w1, xform1);
-
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pmulhw(__m64 __m1, __m64 __m2) {
-  return _mm_mulhi_pi16(__m1, __m2);
-}
-
-/* Multiply four 16-bit values in M1 by four 16-bit values in M2 and produce
-   the low 16 bits of the results.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_mullo_pi16(__m64 __m1, __m64 __m2) {
-  __vector signed short a, b, c;
-
-  a = (__vector signed short)vec_splats(__m1);
-  b = (__vector signed short)vec_splats(__m2);
-  c = a * b;
-  return (__m64)((__vector long long)c)[0];
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pmullw(__m64 __m1, __m64 __m2) {
-  return _mm_mullo_pi16(__m1, __m2);
-}
-
-/* Shift four 16-bit values in M left by COUNT.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sll_pi16(__m64 __m, __m64 __count) {
-  __vector signed short m, r;
-  __vector unsigned short c;
-
-  if (__count <= 15) {
-    m = (__vector signed short)vec_splats(__m);
-    c = (__vector unsigned short)vec_splats((unsigned short)__count);
-    r = vec_sl(m, (__vector unsigned short)c);
-    return (__m64)((__vector long long)r)[0];
-  } else
-    return (0);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psllw(__m64 __m, __m64 __count) {
-  return _mm_sll_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_slli_pi16(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_sll_pi16.  */
-  return _mm_sll_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psllwi(__m64 __m, int __count) {
-  return _mm_slli_pi16(__m, __count);
-}
-
-/* Shift two 32-bit values in M left by COUNT.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sll_pi32(__m64 __m, __m64 __count) {
-  __m64_union m, res;
-
-  m.as_m64 = __m;
-
-  res.as_int[0] = m.as_int[0] << __count;
-  res.as_int[1] = m.as_int[1] << __count;
-  return (res.as_m64);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pslld(__m64 __m, __m64 __count) {
-  return _mm_sll_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_slli_pi32(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_sll_pi32.  */
-  return _mm_sll_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_pslldi(__m64 __m, int __count) {
-  return _mm_slli_pi32(__m, __count);
-}
-
-/* Shift four 16-bit values in M right by COUNT; shift in the sign bit.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sra_pi16(__m64 __m, __m64 __count) {
-  __vector signed short m, r;
-  __vector unsigned short c;
-
-  if (__count <= 15) {
-    m = (__vector signed short)vec_splats(__m);
-    c = (__vector unsigned short)vec_splats((unsigned short)__count);
-    r = vec_sra(m, (__vector unsigned short)c);
-    return (__m64)((__vector long long)r)[0];
-  } else
-    return (0);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psraw(__m64 __m, __m64 __count) {
-  return _mm_sra_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srai_pi16(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_sra_pi32.  */
-  return _mm_sra_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrawi(__m64 __m, int __count) {
-  return _mm_srai_pi16(__m, __count);
-}
-
-/* Shift two 32-bit values in M right by COUNT; shift in the sign bit.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_sra_pi32(__m64 __m, __m64 __count) {
-  __m64_union m, res;
-
-  m.as_m64 = __m;
-
-  res.as_int[0] = m.as_int[0] >> __count;
-  res.as_int[1] = m.as_int[1] >> __count;
-  return (res.as_m64);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrad(__m64 __m, __m64 __count) {
-  return _mm_sra_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srai_pi32(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_sra_pi32.  */
-  return _mm_sra_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psradi(__m64 __m, int __count) {
-  return _mm_srai_pi32(__m, __count);
-}
-
-/* Shift four 16-bit values in M right by COUNT; shift in zeros.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srl_pi16(__m64 __m, __m64 __count) {
-  __vector unsigned short m, r;
-  __vector unsigned short c;
-
-  if (__count <= 15) {
-    m = (__vector unsigned short)vec_splats(__m);
-    c = (__vector unsigned short)vec_splats((unsigned short)__count);
-    r = vec_sr(m, (__vector unsigned short)c);
-    return (__m64)((__vector long long)r)[0];
-  } else
-    return (0);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrlw(__m64 __m, __m64 __count) {
-  return _mm_srl_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srli_pi16(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_sra_pi32.  */
-  return _mm_srl_pi16(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrlwi(__m64 __m, int __count) {
-  return _mm_srli_pi16(__m, __count);
-}
-
-/* Shift two 32-bit values in M right by COUNT; shift in zeros.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srl_pi32(__m64 __m, __m64 __count) {
-  __m64_union m, res;
-
-  m.as_m64 = __m;
-
-  res.as_int[0] = (unsigned int)m.as_int[0] >> __count;
-  res.as_int[1] = (unsigned int)m.as_int[1] >> __count;
-  return (res.as_m64);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrld(__m64 __m, __m64 __count) {
-  return _mm_srl_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_srli_pi32(__m64 __m, int __count) {
-  /* Promote int to long then invoke mm_srl_pi32.  */
-  return _mm_srl_pi32(__m, __count);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _m_psrldi(__m64 __m, int __count) {
-  return _mm_srli_pi32(__m, __count);
-}
-#endif /* _ARCH_PWR8 */
-
-/* Creates a vector of two 32-bit values; I0 is least significant.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set_pi32(int __i1, int __i0) {
-  __m64_union res;
-
-  res.as_int[0] = __i0;
-  res.as_int[1] = __i1;
-  return (res.as_m64);
-}
-
-/* Creates a vector of four 16-bit values; W0 is least significant.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set_pi16(short __w3, short __w2, short __w1, short __w0) {
-  __m64_union res;
-
-  res.as_short[0] = __w0;
-  res.as_short[1] = __w1;
-  res.as_short[2] = __w2;
-  res.as_short[3] = __w3;
-  return (res.as_m64);
-}
-
-/* Creates a vector of eight 8-bit values; B0 is least significant.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set_pi8(char __b7, char __b6, char __b5, char __b4, char __b3,
-                char __b2, char __b1, char __b0) {
-  __m64_union res;
-
-  res.as_char[0] = __b0;
-  res.as_char[1] = __b1;
-  res.as_char[2] = __b2;
-  res.as_char[3] = __b3;
-  res.as_char[4] = __b4;
-  res.as_char[5] = __b5;
-  res.as_char[6] = __b6;
-  res.as_char[7] = __b7;
-  return (res.as_m64);
-}
-
-/* Similar, but with the arguments in reverse order.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_setr_pi32(int __i0, int __i1) {
-  __m64_union res;
-
-  res.as_int[0] = __i0;
-  res.as_int[1] = __i1;
-  return (res.as_m64);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_setr_pi16(short __w0, short __w1, short __w2, short __w3) {
-  return _mm_set_pi16(__w3, __w2, __w1, __w0);
-}
-
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_setr_pi8(char __b0, char __b1, char __b2, char __b3, char __b4,
-                 char __b5, char __b6, char __b7) {
-  return _mm_set_pi8(__b7, __b6, __b5, __b4, __b3, __b2, __b1, __b0);
-}
-
-/* Creates a vector of two 32-bit values, both elements containing I.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set1_pi32(int __i) {
-  __m64_union res;
-
-  res.as_int[0] = __i;
-  res.as_int[1] = __i;
-  return (res.as_m64);
-}
-
-/* Creates a vector of four 16-bit values, all elements containing W.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set1_pi16(short __w) {
-#if _ARCH_PWR9
-  __vector signed short w;
-
-  w = (__vector signed short)vec_splats(__w);
-  return (__m64)((__vector long long)w)[0];
-#else
-  __m64_union res;
-
-  res.as_short[0] = __w;
-  res.as_short[1] = __w;
-  res.as_short[2] = __w;
-  res.as_short[3] = __w;
-  return (res.as_m64);
-#endif
-}
-
-/* Creates a vector of eight 8-bit values, all elements containing B.  */
-extern __inline __m64
-    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-    _mm_set1_pi8(signed char __b) {
-#if _ARCH_PWR8
-  __vector signed char b;
-
-  b = (__vector signed char)vec_splats(__b);
-  return (__m64)((__vector long long)b)[0];
-#else
-  __m64_union res;
-
-  res.as_char[0] = __b;
-  res.as_char[1] = __b;
-  res.as_char[2] = __b;
-  res.as_char[3] = __b;
-  res.as_char[4] = __b;
-  res.as_char[5] = __b;
-  res.as_char[6] = __b;
-  res.as_char[7] = __b;
-  return (res.as_m64);
-#endif
-}
-#endif /* _MMINTRIN_H_INCLUDED */
diff --git a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h b/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
deleted file mode 100644
index 1b322b6..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/ppc_wrappers/xmmintrin.h
+++ /dev/null
@@ -1,1838 +0,0 @@
-/*===---- xmmintrin.h - Implementation of SSE intrinsics on PowerPC --------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-/* Implemented from the specification included in the Intel C++ Compiler
-   User Guide and Reference, version 9.0.  */
-
-#ifndef NO_WARN_X86_INTRINSICS
-/* This header file is to help porting code using Intel intrinsics
-   explicitly from x86_64 to powerpc64/powerpc64le.
-
-   Since X86 SSE intrinsics mainly handles __m128 type, PowerPC
-   VMX/VSX ISA is a good match for vector float SIMD operations.
-   However scalar float operations in vector (XMM) registers require
-   the POWER8 VSX ISA (2.07) level. There are differences for data
-   format and placement of float scalars in the vector register, which
-   require extra steps to match SSE scalar float semantics on POWER.
-
-   It should be noted that there's much difference between X86_64's
-   MXSCR and PowerISA's FPSCR/VSCR registers. It's recommended to use
-   portable <fenv.h> instead of access MXSCR directly.
-
-   Most SSE scalar float intrinsic operations can be performed more
-   efficiently as C language float scalar operations or optimized to
-   use vector SIMD operations. We recommend this for new applications. */
-#error "Please read comment above.  Use -DNO_WARN_X86_INTRINSICS to disable this error."
-#endif
-
-#ifndef _XMMINTRIN_H_INCLUDED
-#define _XMMINTRIN_H_INCLUDED
-
-/* Define four value permute mask */
-#define _MM_SHUFFLE(w,x,y,z) (((w) << 6) | ((x) << 4) | ((y) << 2) | (z))
-
-#include <altivec.h>
-
-/* Avoid collisions between altivec.h and strict adherence to C++ and
-   C11 standards.  This should eventually be done inside altivec.h itself,
-   but only after testing a full distro build.  */
-#if defined(__STRICT_ANSI__) && (defined(__cplusplus) || \
-				 (defined(__STDC_VERSION__) &&	\
-				  __STDC_VERSION__ >= 201112L))
-#undef vector
-#undef pixel
-#undef bool
-#endif
-
-/* We need type definitions from the MMX header file.  */
-#include <mmintrin.h>
-
-/* Get _mm_malloc () and _mm_free ().  */
-#if __STDC_HOSTED__
-#include <mm_malloc.h>
-#endif
-
-/* The Intel API is flexible enough that we must allow aliasing with other
-   vector types, and their scalar components.  */
-typedef float __m128 __attribute__ ((__vector_size__ (16), __may_alias__));
-
-/* Unaligned version of the same type.  */
-typedef float __m128_u __attribute__ ((__vector_size__ (16), __may_alias__,
-				       __aligned__ (1)));
-
-/* Internal data types for implementing the intrinsics.  */
-typedef float __v4sf __attribute__ ((__vector_size__ (16)));
-
-/* Create an undefined vector.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_undefined_ps (void)
-{
-  __m128 __Y = __Y;
-  return __Y;
-}
-
-/* Create a vector of zeros.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setzero_ps (void)
-{
-  return __extension__ (__m128){ 0.0f, 0.0f, 0.0f, 0.0f };
-}
-
-/* Load four SPFP values from P.  The address must be 16-byte aligned.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_ps (float const *__P)
-{
-  return ((__m128)vec_ld(0, (__v4sf*)__P));
-}
-
-/* Load four SPFP values from P.  The address need not be 16-byte aligned.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadu_ps (float const *__P)
-{
-  return (vec_vsx_ld(0, __P));
-}
-
-/* Load four SPFP values in reverse order.  The address must be aligned.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadr_ps (float const *__P)
-{
-  __v4sf   __tmp;
-  __m128 result;
-  static const __vector unsigned char permute_vector =
-    { 0x1C, 0x1D, 0x1E, 0x1F, 0x18, 0x19, 0x1A, 0x1B, 0x14, 0x15, 0x16,
-	0x17, 0x10, 0x11, 0x12, 0x13 };
-
-  __tmp = vec_ld (0, (__v4sf *) __P);
-  result = (__m128) vec_perm (__tmp, __tmp, permute_vector);
-  return result;
-}
-
-/* Create a vector with all four elements equal to F.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set1_ps (float __F)
-{
-  return __extension__ (__m128)(__v4sf){ __F, __F, __F, __F };
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_ps1 (float __F)
-{
-  return _mm_set1_ps (__F);
-}
-
-/* Create the vector [Z Y X W].  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_ps (const float __Z, const float __Y, const float __X, const float __W)
-{
-  return __extension__ (__m128)(__v4sf){ __W, __X, __Y, __Z };
-}
-
-/* Create the vector [W X Y Z].  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_setr_ps (float __Z, float __Y, float __X, float __W)
-{
-  return __extension__ (__m128)(__v4sf){ __Z, __Y, __X, __W };
-}
-
-/* Store four SPFP values.  The address must be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_ps (float *__P, __m128 __A)
-{
-  vec_st((__v4sf)__A, 0, (__v4sf*)__P);
-}
-
-/* Store four SPFP values.  The address need not be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storeu_ps (float *__P, __m128 __A)
-{
-  *(__m128_u *)__P = __A;
-}
-
-/* Store four SPFP values in reverse order.  The address must be aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storer_ps (float *__P, __m128 __A)
-{
-  __v4sf   __tmp;
-  static const __vector unsigned char permute_vector =
-    { 0x1C, 0x1D, 0x1E, 0x1F, 0x18, 0x19, 0x1A, 0x1B, 0x14, 0x15, 0x16,
-	0x17, 0x10, 0x11, 0x12, 0x13 };
-
-  __tmp = (__m128) vec_perm (__A, __A, permute_vector);
-
-  _mm_store_ps (__P, __tmp);
-}
-
-/* Store the lower SPFP value across four words.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store1_ps (float *__P, __m128 __A)
-{
-  __v4sf __va = vec_splat((__v4sf)__A, 0);
-  _mm_store_ps (__P, __va);
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_ps1 (float *__P, __m128 __A)
-{
-  _mm_store1_ps (__P, __A);
-}
-
-/* Create a vector with element 0 as F and the rest zero.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_set_ss (float __F)
-{
-  return __extension__ (__m128)(__v4sf){ __F, 0.0f, 0.0f, 0.0f };
-}
-
-/* Sets the low SPFP value of A from the low value of B.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_move_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-
-  return (vec_sel ((__v4sf)__A, (__v4sf)__B, mask));
-}
-
-/* Create a vector with element 0 as *P and the rest zero.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_ss (float const *__P)
-{
-  return _mm_set_ss (*__P);
-}
-
-/* Stores the lower SPFP value.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_store_ss (float *__P, __m128 __A)
-{
-  *__P = ((__v4sf)__A)[0];
-}
-
-/* Perform the respective operation on the lower SPFP (single-precision
-   floating-point) values of A and B; the upper three SPFP values are
-   passed through from A.  */
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_ss (__m128 __A, __m128 __B)
-{
-#ifdef _ARCH_PWR7
-  __m128 a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-     results. So to insure we don't generate spurious exceptions
-     (from the upper double values) we splat the lower double
-     before we to the operation.  */
-  a = vec_splat (__A, 0);
-  b = vec_splat (__B, 0);
-  c = a + b;
-  /* Then we merge the lower float result with the original upper
-     float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-#else
-  __A[0] = __A[0] + __B[0];
-  return (__A);
-#endif
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_ss (__m128 __A, __m128 __B)
-{
-#ifdef _ARCH_PWR7
-  __m128 a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-     results. So to insure we don't generate spurious exceptions
-     (from the upper double values) we splat the lower double
-     before we to the operation.  */
-  a = vec_splat (__A, 0);
-  b = vec_splat (__B, 0);
-  c = a - b;
-  /* Then we merge the lower float result with the original upper
-     float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-#else
-  __A[0] = __A[0] - __B[0];
-  return (__A);
-#endif
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_ss (__m128 __A, __m128 __B)
-{
-#ifdef _ARCH_PWR7
-  __m128 a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-     results. So to insure we don't generate spurious exceptions
-     (from the upper double values) we splat the lower double
-     before we to the operation.  */
-  a = vec_splat (__A, 0);
-  b = vec_splat (__B, 0);
-  c = a * b;
-  /* Then we merge the lower float result with the original upper
-     float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-#else
-  __A[0] = __A[0] * __B[0];
-  return (__A);
-#endif
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_div_ss (__m128 __A, __m128 __B)
-{
-#ifdef _ARCH_PWR7
-  __m128 a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-     results. So to insure we don't generate spurious exceptions
-     (from the upper double values) we splat the lower double
-     before we to the operation.  */
-  a = vec_splat (__A, 0);
-  b = vec_splat (__B, 0);
-  c = a / b;
-  /* Then we merge the lower float result with the original upper
-     float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-#else
-  __A[0] = __A[0] / __B[0];
-  return (__A);
-#endif
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sqrt_ss (__m128 __A)
-{
-  __m128 a, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper double values) we splat the lower double
-   * before we to the operation. */
-  a = vec_splat (__A, 0);
-  c = vec_sqrt (a);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-}
-
-/* Perform the respective operation on the four SPFP values in A and B.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_add_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) ((__v4sf)__A + (__v4sf)__B);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sub_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) ((__v4sf)__A - (__v4sf)__B);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mul_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) ((__v4sf)__A * (__v4sf)__B);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_div_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) ((__v4sf)__A / (__v4sf)__B);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sqrt_ps (__m128 __A)
-{
-  return (vec_sqrt ((__v4sf)__A));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_rcp_ps (__m128 __A)
-{
-  return (vec_re ((__v4sf)__A));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_rsqrt_ps (__m128 __A)
-{
-  return (vec_rsqrte (__A));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_rcp_ss (__m128 __A)
-{
-  __m128 a, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper double values) we splat the lower double
-   * before we to the operation. */
-  a = vec_splat (__A, 0);
-  c = _mm_rcp_ps (a);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_rsqrt_ss (__m128 __A)
-{
-  __m128 a, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower double)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper double values) we splat the lower double
-   * before we to the operation. */
-  a = vec_splat (__A, 0);
-  c = vec_rsqrte (a);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return (vec_sel (__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_ss (__m128 __A, __m128 __B)
-{
-  __v4sf a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower float)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper float values) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf)__A, 0);
-  b = vec_splat ((__v4sf)__B, 0);
-  c = vec_min (a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return (vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_ss (__m128 __A, __m128 __B)
-{
-  __v4sf a, b, c;
-  static const __vector unsigned int mask = {0xffffffff, 0, 0, 0};
-  /* PowerISA VSX does not allow partial (for just lower float)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper float values) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat (__A, 0);
-  b = vec_splat (__B, 0);
-  c = vec_max (a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return (vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_ps (__m128 __A, __m128 __B)
-{
-  __vector __bool int m = vec_cmpgt ((__v4sf) __B, (__v4sf) __A);
-  return vec_sel (__B, __A, m);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_ps (__m128 __A, __m128 __B)
-{
-  __vector __bool int m = vec_cmpgt ((__v4sf) __A, (__v4sf) __B);
-  return vec_sel (__B, __A, m);
-}
-
-/* Perform logical bit-wise operations on 128-bit values.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_and_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_and ((__v4sf)__A, (__v4sf)__B));
-//  return __builtin_ia32_andps (__A, __B);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_andnot_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_andc ((__v4sf)__B, (__v4sf)__A));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_or_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_or ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_xor_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_xor ((__v4sf)__A, (__v4sf)__B));
-}
-
-/* Perform a comparison on the four SPFP values of A and B.  For each
-   element, if the comparison is true, place a mask of all ones in the
-   result, otherwise a mask of zeros.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmpeq ((__v4sf)__A,(__v4sf) __B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmplt ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmple_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmple ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmpgt ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpge_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmpge ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline  __m128  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpneq_ps (__m128  __A, __m128  __B)
-{
-  __v4sf temp = (__v4sf ) vec_cmpeq ((__v4sf) __A, (__v4sf)__B);
-  return ((__m128)vec_nor (temp, temp));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnlt_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmpge ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnle_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmpgt ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpngt_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmple ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnge_ps (__m128 __A, __m128 __B)
-{
-  return ((__m128)vec_cmplt ((__v4sf)__A, (__v4sf)__B));
-}
-
-extern __inline  __m128  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpord_ps (__m128  __A, __m128  __B)
-{
-  __vector unsigned int a, b;
-  __vector unsigned int c, d;
-  static const __vector unsigned int float_exp_mask =
-    { 0x7f800000, 0x7f800000, 0x7f800000, 0x7f800000 };
-
-  a = (__vector unsigned int) vec_abs ((__v4sf)__A);
-  b = (__vector unsigned int) vec_abs ((__v4sf)__B);
-  c = (__vector unsigned int) vec_cmpgt (float_exp_mask, a);
-  d = (__vector unsigned int) vec_cmpgt (float_exp_mask, b);
-  return ((__m128 ) vec_and (c, d));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpunord_ps (__m128 __A, __m128 __B)
-{
-  __vector unsigned int a, b;
-  __vector unsigned int c, d;
-  static const __vector unsigned int float_exp_mask =
-    { 0x7f800000, 0x7f800000, 0x7f800000, 0x7f800000 };
-
-  a = (__vector unsigned int) vec_abs ((__v4sf)__A);
-  b = (__vector unsigned int) vec_abs ((__v4sf)__B);
-  c = (__vector unsigned int) vec_cmpgt (a, float_exp_mask);
-  d = (__vector unsigned int) vec_cmpgt (b, float_exp_mask);
-  return ((__m128 ) vec_or (c, d));
-}
-
-/* Perform a comparison on the lower SPFP values of A and B.  If the
-   comparison is true, place a mask of all ones in the result, otherwise a
-   mask of zeros.  The upper three SPFP values are passed through from A.  */
-extern __inline  __m128  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpeq_ss (__m128  __A, __m128  __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpeq(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmplt_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmplt(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmple_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmple(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpgt_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpgt(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpge_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpge(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpneq_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpeq(a, b);
-  c = vec_nor (c, c);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnlt_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpge(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnle_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmpgt(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpngt_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we to the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmple(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpnge_ss (__m128 __A, __m128 __B)
-{
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-  __v4sf a, b, c;
-  /* PowerISA VMX does not allow partial (for just element 0)
-   * results. So to insure we don't generate spurious exceptions
-   * (from the upper elements) we splat the lower float
-   * before we do the operation. */
-  a = vec_splat ((__v4sf) __A, 0);
-  b = vec_splat ((__v4sf) __B, 0);
-  c = (__v4sf) vec_cmplt(a, b);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpord_ss (__m128 __A, __m128 __B)
-{
-  __vector unsigned int a, b;
-  __vector unsigned int c, d;
-  static const __vector unsigned int float_exp_mask =
-    { 0x7f800000, 0x7f800000, 0x7f800000, 0x7f800000 };
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-
-  a = (__vector unsigned int) vec_abs ((__v4sf)__A);
-  b = (__vector unsigned int) vec_abs ((__v4sf)__B);
-  c = (__vector unsigned int) vec_cmpgt (float_exp_mask, a);
-  d = (__vector unsigned int) vec_cmpgt (float_exp_mask, b);
-  c = vec_and (c, d);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, (__v4sf)c, mask));
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cmpunord_ss (__m128 __A, __m128 __B)
-{
-  __vector unsigned int a, b;
-  __vector unsigned int c, d;
-  static const __vector unsigned int float_exp_mask =
-    { 0x7f800000, 0x7f800000, 0x7f800000, 0x7f800000 };
-  static const __vector unsigned int mask =
-    { 0xffffffff, 0, 0, 0 };
-
-  a = (__vector unsigned int) vec_abs ((__v4sf)__A);
-  b = (__vector unsigned int) vec_abs ((__v4sf)__B);
-  c = (__vector unsigned int) vec_cmpgt (a, float_exp_mask);
-  d = (__vector unsigned int) vec_cmpgt (b, float_exp_mask);
-  c = vec_or (c, d);
-  /* Then we merge the lower float result with the original upper
-   * float elements from __A.  */
-  return ((__m128)vec_sel ((__v4sf)__A, (__v4sf)c, mask));
-}
-
-/* Compare the lower SPFP values of A and B and return 1 if true
-   and 0 if false.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comieq_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] == __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comilt_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] < __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comile_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] <= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comigt_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] > __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comige_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] >= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_comineq_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] != __B[0]);
-}
-
-/* FIXME
- * The __mm_ucomi??_ss implementations below are exactly the same as
- * __mm_comi??_ss because GCC for PowerPC only generates unordered
- * compares (scalar and vector).
- * Technically __mm_comieq_ss et al should be using the ordered
- * compare and signal for QNaNs.
- * The __mm_ucomieq_sd et all should be OK, as is.
- */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomieq_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] == __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomilt_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] < __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomile_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] <= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomigt_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] > __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomige_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] >= __B[0]);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_ucomineq_ss (__m128 __A, __m128 __B)
-{
-  return (__A[0] != __B[0]);
-}
-
-extern __inline float __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtss_f32 (__m128 __A)
-{
-  return ((__v4sf)__A)[0];
-}
-
-/* Convert the lower SPFP value to a 32-bit integer according to the current
-   rounding mode.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtss_si32 (__m128 __A)
-{
-  __m64 res = 0;
-#ifdef _ARCH_PWR8
-  double dtmp;
-  __asm__(
-#ifdef __LITTLE_ENDIAN__
-      "xxsldwi %x0,%x0,%x0,3;\n"
-#endif
-      "xscvspdp %x2,%x0;\n"
-      "fctiw  %2,%2;\n"
-      "mfvsrd  %1,%x2;\n"
-      : "+wa" (__A),
-        "=r" (res),
-        "=f" (dtmp)
-      : );
-#else
-  res = __builtin_rint(__A[0]);
-#endif
-  return (res);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvt_ss2si (__m128 __A)
-{
-  return _mm_cvtss_si32 (__A);
-}
-
-/* Convert the lower SPFP value to a 32-bit integer according to the
-   current rounding mode.  */
-
-/* Intel intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtss_si64 (__m128 __A)
-{
-  __m64 res = 0;
-#ifdef _ARCH_PWR8
-  double dtmp;
-  __asm__(
-#ifdef __LITTLE_ENDIAN__
-      "xxsldwi %x0,%x0,%x0,3;\n"
-#endif
-      "xscvspdp %x2,%x0;\n"
-      "fctid  %2,%2;\n"
-      "mfvsrd  %1,%x2;\n"
-      : "+wa" (__A),
-        "=r" (res),
-        "=f" (dtmp)
-      : );
-#else
-  res = __builtin_llrint(__A[0]);
-#endif
-  return (res);
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtss_si64x (__m128 __A)
-{
-  return _mm_cvtss_si64 ((__v4sf) __A);
-}
-
-/* Constants for use with _mm_prefetch.  */
-enum _mm_hint
-{
-  /* _MM_HINT_ET is _MM_HINT_T with set 3rd bit.  */
-  _MM_HINT_ET0 = 7,
-  _MM_HINT_ET1 = 6,
-  _MM_HINT_T0 = 3,
-  _MM_HINT_T1 = 2,
-  _MM_HINT_T2 = 1,
-  _MM_HINT_NTA = 0
-};
-
-/* Loads one cache line from address P to a location "closer" to the
-   processor.  The selector I specifies the type of prefetch operation.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_prefetch (const void *__P, enum _mm_hint __I)
-{
-  /* Current PowerPC will ignores the hint parameters.  */
-  __builtin_prefetch (__P);
-}
-
-/* Convert the two lower SPFP values to 32-bit integers according to the
-   current rounding mode.  Return the integers in packed form.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtps_pi32 (__m128 __A)
-{
-  /* Splat two lower SPFP values to both halves.  */
-  __v4sf temp, rounded;
-  __vector unsigned long long result;
-
-  /* Splat two lower SPFP values to both halves.  */
-  temp = (__v4sf) vec_splat ((__vector long long)__A, 0);
-  rounded = vec_rint(temp);
-  result = (__vector unsigned long long) vec_cts (rounded, 0);
-
-  return (__m64) ((__vector long long) result)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvt_ps2pi (__m128 __A)
-{
-  return _mm_cvtps_pi32 (__A);
-}
-
-/* Truncate the lower SPFP value to a 32-bit integer.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttss_si32 (__m128 __A)
-{
-  /* Extract the lower float element.  */
-  float temp = __A[0];
-  /* truncate to 32-bit integer and return.  */
-  return temp;
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtt_ss2si (__m128 __A)
-{
-  return _mm_cvttss_si32 (__A);
-}
-
-/* Intel intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttss_si64 (__m128 __A)
-{
-  /* Extract the lower float element.  */
-  float temp = __A[0];
-  /* truncate to 32-bit integer and return.  */
-  return temp;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttss_si64x (__m128 __A)
-{
-  /* Extract the lower float element.  */
-  float temp = __A[0];
-  /* truncate to 32-bit integer and return.  */
-  return temp;
-}
-
-/* Truncate the two lower SPFP values to 32-bit integers.  Return the
-   integers in packed form.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvttps_pi32 (__m128 __A)
-{
-  __v4sf temp;
-  __vector unsigned long long result;
-
-  /* Splat two lower SPFP values to both halves.  */
-  temp = (__v4sf) vec_splat ((__vector long long)__A, 0);
-  result = (__vector unsigned long long) vec_cts (temp, 0);
-
-  return (__m64) ((__vector long long) result)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtt_ps2pi (__m128 __A)
-{
-  return _mm_cvttps_pi32 (__A);
-}
-
-/* Convert B to a SPFP value and insert it as element zero in A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi32_ss (__m128 __A, int __B)
-{
-  float temp = __B;
-  __A[0] = temp;
-
-  return __A;
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvt_si2ss (__m128 __A, int __B)
-{
-  return _mm_cvtsi32_ss (__A, __B);
-}
-
-/* Convert B to a SPFP value and insert it as element zero in A.  */
-/* Intel intrinsic.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64_ss (__m128 __A, long long __B)
-{
-  float temp = __B;
-  __A[0] = temp;
-
-  return __A;
-}
-
-/* Microsoft intrinsic.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtsi64x_ss (__m128 __A, long long __B)
-{
-  return _mm_cvtsi64_ss (__A, __B);
-}
-
-/* Convert the two 32-bit values in B to SPFP form and insert them
-   as the two lower elements in A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpi32_ps (__m128        __A, __m64        __B)
-{
-  __vector signed int vm1;
-  __vector float vf1;
-
-  vm1 = (__vector signed int) (__vector unsigned long long) {__B, __B};
-  vf1 = (__vector float) vec_ctf (vm1, 0);
-
-  return ((__m128) (__vector unsigned long long)
-    { ((__vector unsigned long long)vf1) [0],
-	((__vector unsigned long long)__A) [1]});
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvt_pi2ps (__m128 __A, __m64 __B)
-{
-  return _mm_cvtpi32_ps (__A, __B);
-}
-
-/* Convert the four signed 16-bit values in A to SPFP form.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpi16_ps (__m64 __A)
-{
-  __vector signed short vs8;
-  __vector signed int vi4;
-  __vector float vf1;
-
-  vs8 = (__vector signed short) (__vector unsigned long long) { __A, __A };
-  vi4 = vec_vupklsh (vs8);
-  vf1 = (__vector float) vec_ctf (vi4, 0);
-
-  return (__m128) vf1;
-}
-
-/* Convert the four unsigned 16-bit values in A to SPFP form.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpu16_ps (__m64 __A)
-{
-  const __vector unsigned short zero =
-    { 0, 0, 0, 0, 0, 0, 0, 0 };
-  __vector unsigned short vs8;
-  __vector unsigned int vi4;
-  __vector float vf1;
-
-  vs8 = (__vector unsigned short) (__vector unsigned long long) { __A, __A };
-  vi4 = (__vector unsigned int) vec_mergel
-#ifdef __LITTLE_ENDIAN__
-                                           (vs8, zero);
-#else
-                                           (zero, vs8);
-#endif
-  vf1 = (__vector float) vec_ctf (vi4, 0);
-
-  return (__m128) vf1;
-}
-
-/* Convert the low four signed 8-bit values in A to SPFP form.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpi8_ps (__m64 __A)
-{
-  __vector signed char vc16;
-  __vector signed short vs8;
-  __vector signed int vi4;
-  __vector float vf1;
-
-  vc16 = (__vector signed char) (__vector unsigned long long) { __A, __A };
-  vs8 = vec_vupkhsb (vc16);
-  vi4 = vec_vupkhsh (vs8);
-  vf1 = (__vector float) vec_ctf (vi4, 0);
-
-  return (__m128) vf1;
-}
-
-/* Convert the low four unsigned 8-bit values in A to SPFP form.  */
-extern __inline  __m128  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-
-_mm_cvtpu8_ps (__m64  __A)
-{
-  const __vector unsigned char zero =
-    { 0, 0, 0, 0, 0, 0, 0, 0 };
-  __vector unsigned char vc16;
-  __vector unsigned short vs8;
-  __vector unsigned int vi4;
-  __vector float vf1;
-
-  vc16 = (__vector unsigned char) (__vector unsigned long long) { __A, __A };
-#ifdef __LITTLE_ENDIAN__
-  vs8 = (__vector unsigned short) vec_mergel (vc16, zero);
-  vi4 = (__vector unsigned int) vec_mergeh (vs8,
-					    (__vector unsigned short) zero);
-#else
-  vs8 = (__vector unsigned short) vec_mergel (zero, vc16);
-  vi4 = (__vector unsigned int) vec_mergeh ((__vector unsigned short) zero,
-                                            vs8);
-#endif
-  vf1 = (__vector float) vec_ctf (vi4, 0);
-
-  return (__m128) vf1;
-}
-
-/* Convert the four signed 32-bit values in A and B to SPFP form.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtpi32x2_ps (__m64 __A, __m64 __B)
-{
-  __vector signed int vi4;
-  __vector float vf4;
-
-  vi4 = (__vector signed int) (__vector unsigned long long) { __A, __B };
-  vf4 = (__vector float) vec_ctf (vi4, 0);
-  return (__m128) vf4;
-}
-
-/* Convert the four SPFP values in A to four signed 16-bit integers.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtps_pi16 (__m128 __A)
-{
-  __v4sf rounded;
-  __vector signed int temp;
-  __vector unsigned long long result;
-
-  rounded = vec_rint(__A);
-  temp = vec_cts (rounded, 0);
-  result = (__vector unsigned long long) vec_pack (temp, temp);
-
-  return (__m64) ((__vector long long) result)[0];
-}
-
-/* Convert the four SPFP values in A to four signed 8-bit integers.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_cvtps_pi8 (__m128 __A)
-{
-  __v4sf rounded;
-  __vector signed int tmp_i;
-  static const __vector signed int zero = {0, 0, 0, 0};
-  __vector signed short tmp_s;
-  __vector signed char res_v;
-
-  rounded = vec_rint(__A);
-  tmp_i = vec_cts (rounded, 0);
-  tmp_s = vec_pack (tmp_i, zero);
-  res_v = vec_pack (tmp_s, tmp_s);
-  return (__m64) ((__vector long long) res_v)[0];
-}
-
-/* Selects four specific SPFP values from A and B based on MASK.  */
-extern __inline  __m128  __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-
-_mm_shuffle_ps (__m128  __A, __m128  __B, int const __mask)
-{
-  unsigned long element_selector_10 = __mask & 0x03;
-  unsigned long element_selector_32 = (__mask >> 2) & 0x03;
-  unsigned long element_selector_54 = (__mask >> 4) & 0x03;
-  unsigned long element_selector_76 = (__mask >> 6) & 0x03;
-  static const unsigned int permute_selectors[4] =
-    {
-#ifdef __LITTLE_ENDIAN__
-      0x03020100, 0x07060504, 0x0B0A0908, 0x0F0E0D0C
-#else
-      0x00010203, 0x04050607, 0x08090A0B, 0x0C0D0E0F
-#endif
-    };
-  __vector unsigned int t;
-
-  t[0] = permute_selectors[element_selector_10];
-  t[1] = permute_selectors[element_selector_32];
-  t[2] = permute_selectors[element_selector_54] + 0x10101010;
-  t[3] = permute_selectors[element_selector_76] + 0x10101010;
-  return vec_perm ((__v4sf) __A, (__v4sf)__B, (__vector unsigned char)t);
-}
-
-/* Selects and interleaves the upper two SPFP values from A and B.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpackhi_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) vec_vmrglw ((__v4sf) __A, (__v4sf)__B);
-}
-
-/* Selects and interleaves the lower two SPFP values from A and B.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_unpacklo_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) vec_vmrghw ((__v4sf) __A, (__v4sf)__B);
-}
-
-/* Sets the upper two SPFP values with 64-bits of data loaded from P;
-   the lower two values are passed through from A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadh_pi (__m128 __A, __m64 const *__P)
-{
-  __vector unsigned long long __a = (__vector unsigned long long)__A;
-  __vector unsigned long long __p = vec_splats(*__P);
-  __a [1] = __p [1];
-
-  return (__m128)__a;
-}
-
-/* Stores the upper two SPFP values of A into P.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storeh_pi (__m64 *__P, __m128 __A)
-{
-  __vector unsigned long long __a = (__vector unsigned long long) __A;
-
-  *__P = __a[1];
-}
-
-/* Moves the upper two values of B into the lower two values of A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movehl_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) vec_mergel ((__vector unsigned long long)__B,
-			      (__vector unsigned long long)__A);
-}
-
-/* Moves the lower two values of B into the upper two values of A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movelh_ps (__m128 __A, __m128 __B)
-{
-  return (__m128) vec_mergeh ((__vector unsigned long long)__A,
-			      (__vector unsigned long long)__B);
-}
-
-/* Sets the lower two SPFP values with 64-bits of data loaded from P;
-   the upper two values are passed through from A.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_loadl_pi (__m128 __A, __m64 const *__P)
-{
-  __vector unsigned long long __a = (__vector unsigned long long)__A;
-  __vector unsigned long long __p = vec_splats(*__P);
-  __a [0] = __p [0];
-
-  return (__m128)__a;
-}
-
-/* Stores the lower two SPFP values of A into P.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_storel_pi (__m64 *__P, __m128 __A)
-{
-  __vector unsigned long long __a = (__vector unsigned long long) __A;
-
-  *__P = __a[0];
-}
-
-#ifdef _ARCH_PWR8
-/* Intrinsic functions that require PowerISA 2.07 minimum.  */
-
-/* Creates a 4-bit mask from the most significant bits of the SPFP values.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movemask_ps (__m128  __A)
-{
-  __vector unsigned long long result;
-  static const __vector unsigned int perm_mask =
-    {
-#ifdef __LITTLE_ENDIAN__
-	0x00204060, 0x80808080, 0x80808080, 0x80808080
-#else
-      0x80808080, 0x80808080, 0x80808080, 0x00204060
-#endif
-    };
-
-  result = ((__vector unsigned long long)
-	    vec_vbpermq ((__vector unsigned char) __A,
-			 (__vector unsigned char) perm_mask));
-
-#ifdef __LITTLE_ENDIAN__
-  return result[1];
-#else
-  return result[0];
-#endif
-}
-#endif /* _ARCH_PWR8 */
-
-/* Create a vector with all four elements equal to *P.  */
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load1_ps (float const *__P)
-{
-  return _mm_set1_ps (*__P);
-}
-
-extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_load_ps1 (float const *__P)
-{
-  return _mm_load1_ps (__P);
-}
-
-/* Extracts one of the four words of A.  The selector N must be immediate.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_extract_pi16 (__m64 const __A, int const __N)
-{
-  unsigned int shiftr = __N & 3;
-#ifdef __BIG_ENDIAN__
-  shiftr = 3 - shiftr;
-#endif
-
-  return ((__A >> (shiftr * 16)) & 0xffff);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pextrw (__m64 const __A, int const __N)
-{
-  return _mm_extract_pi16 (__A, __N);
-}
-
-/* Inserts word D into one of four words of A.  The selector N must be
-   immediate.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_insert_pi16 (__m64 const __A, int const __D, int const __N)
-{
-  const int shiftl = (__N & 3) * 16;
-  const __m64 shiftD = (const __m64) __D << shiftl;
-  const __m64 mask = 0xffffUL << shiftl;
-  __m64 result = (__A & (~mask)) | (shiftD & mask);
-
-  return (result);
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pinsrw (__m64 const __A, int const __D, int const __N)
-{
-  return _mm_insert_pi16 (__A, __D, __N);
-}
-
-/* Compute the element-wise maximum of signed 16-bit values.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-
-_mm_max_pi16 (__m64 __A, __m64 __B)
-{
-#if _ARCH_PWR8
-  __vector signed short a, b, r;
-  __vector __bool short c;
-
-  a = (__vector signed short)vec_splats (__A);
-  b = (__vector signed short)vec_splats (__B);
-  c = (__vector __bool short)vec_cmpgt (a, b);
-  r = vec_sel (b, a, c);
-  return (__m64) ((__vector long long) r)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __A;
-  m2.as_m64 = __B;
-
-  res.as_short[0] =
-      (m1.as_short[0] > m2.as_short[0]) ? m1.as_short[0] : m2.as_short[0];
-  res.as_short[1] =
-      (m1.as_short[1] > m2.as_short[1]) ? m1.as_short[1] : m2.as_short[1];
-  res.as_short[2] =
-      (m1.as_short[2] > m2.as_short[2]) ? m1.as_short[2] : m2.as_short[2];
-  res.as_short[3] =
-      (m1.as_short[3] > m2.as_short[3]) ? m1.as_short[3] : m2.as_short[3];
-
-  return (__m64) res.as_m64;
-#endif
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pmaxsw (__m64 __A, __m64 __B)
-{
-  return _mm_max_pi16 (__A, __B);
-}
-
-/* Compute the element-wise maximum of unsigned 8-bit values.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_max_pu8 (__m64 __A, __m64 __B)
-{
-#if _ARCH_PWR8
-  __vector unsigned char a, b, r;
-  __vector __bool char c;
-
-  a = (__vector unsigned char)vec_splats (__A);
-  b = (__vector unsigned char)vec_splats (__B);
-  c = (__vector __bool char)vec_cmpgt (a, b);
-  r = vec_sel (b, a, c);
-  return (__m64) ((__vector long long) r)[0];
-#else
-  __m64_union m1, m2, res;
-  long i;
-
-  m1.as_m64 = __A;
-  m2.as_m64 = __B;
-
-
-  for (i = 0; i < 8; i++)
-  res.as_char[i] =
-      ((unsigned char) m1.as_char[i] > (unsigned char) m2.as_char[i]) ?
-	  m1.as_char[i] : m2.as_char[i];
-
-  return (__m64) res.as_m64;
-#endif
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pmaxub (__m64 __A, __m64 __B)
-{
-  return _mm_max_pu8 (__A, __B);
-}
-
-/* Compute the element-wise minimum of signed 16-bit values.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_pi16 (__m64 __A, __m64 __B)
-{
-#if _ARCH_PWR8
-  __vector signed short a, b, r;
-  __vector __bool short c;
-
-  a = (__vector signed short)vec_splats (__A);
-  b = (__vector signed short)vec_splats (__B);
-  c = (__vector __bool short)vec_cmplt (a, b);
-  r = vec_sel (b, a, c);
-  return (__m64) ((__vector long long) r)[0];
-#else
-  __m64_union m1, m2, res;
-
-  m1.as_m64 = __A;
-  m2.as_m64 = __B;
-
-  res.as_short[0] =
-      (m1.as_short[0] < m2.as_short[0]) ? m1.as_short[0] : m2.as_short[0];
-  res.as_short[1] =
-      (m1.as_short[1] < m2.as_short[1]) ? m1.as_short[1] : m2.as_short[1];
-  res.as_short[2] =
-      (m1.as_short[2] < m2.as_short[2]) ? m1.as_short[2] : m2.as_short[2];
-  res.as_short[3] =
-      (m1.as_short[3] < m2.as_short[3]) ? m1.as_short[3] : m2.as_short[3];
-
-  return (__m64) res.as_m64;
-#endif
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pminsw (__m64 __A, __m64 __B)
-{
-  return _mm_min_pi16 (__A, __B);
-}
-
-/* Compute the element-wise minimum of unsigned 8-bit values.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_min_pu8 (__m64 __A, __m64 __B)
-{
-#if _ARCH_PWR8
-  __vector unsigned char a, b, r;
-  __vector __bool char c;
-
-  a = (__vector unsigned char)vec_splats (__A);
-  b = (__vector unsigned char)vec_splats (__B);
-  c = (__vector __bool char)vec_cmplt (a, b);
-  r = vec_sel (b, a, c);
-  return (__m64) ((__vector long long) r)[0];
-#else
-  __m64_union m1, m2, res;
-  long i;
-
-  m1.as_m64 = __A;
-  m2.as_m64 = __B;
-
-
-  for (i = 0; i < 8; i++)
-  res.as_char[i] =
-      ((unsigned char) m1.as_char[i] < (unsigned char) m2.as_char[i]) ?
-	  m1.as_char[i] : m2.as_char[i];
-
-  return (__m64) res.as_m64;
-#endif
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pminub (__m64 __A, __m64 __B)
-{
-  return _mm_min_pu8 (__A, __B);
-}
-
-/* Create an 8-bit mask of the signs of 8-bit values.  */
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_movemask_pi8 (__m64 __A)
-{
-  unsigned long long p =
-#ifdef __LITTLE_ENDIAN__
-                         0x0008101820283038UL; // permute control for sign bits
-#else
-                         0x3830282018100800UL; // permute control for sign bits
-#endif
-  return __builtin_bpermd (p, __A);
-}
-
-extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pmovmskb (__m64 __A)
-{
-  return _mm_movemask_pi8 (__A);
-}
-
-/* Multiply four unsigned 16-bit values in A by four unsigned 16-bit values
-   in B and produce the high 16 bits of the 32-bit results.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_mulhi_pu16 (__m64 __A, __m64 __B)
-{
-  __vector unsigned short a, b;
-  __vector unsigned short c;
-  __vector unsigned int w0, w1;
-  __vector unsigned char xform1 = {
-#ifdef __LITTLE_ENDIAN__
-      0x02, 0x03, 0x12, 0x13,  0x06, 0x07, 0x16, 0x17,
-      0x0A, 0x0B, 0x1A, 0x1B,  0x0E, 0x0F, 0x1E, 0x1F
-#else
-      0x00, 0x01, 0x10, 0x11,  0x04, 0x05, 0x14, 0x15,
-      0x00, 0x01, 0x10, 0x11,  0x04, 0x05, 0x14, 0x15
-#endif
-    };
-
-  a = (__vector unsigned short)vec_splats (__A);
-  b = (__vector unsigned short)vec_splats (__B);
-
-  w0 = vec_vmuleuh (a, b);
-  w1 = vec_vmulouh (a, b);
-  c = (__vector unsigned short)vec_perm (w0, w1, xform1);
-
-  return (__m64) ((__vector long long) c)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pmulhuw (__m64 __A, __m64 __B)
-{
-  return _mm_mulhi_pu16 (__A, __B);
-}
-
-/* Return a combination of the four 16-bit values in A.  The selector
-   must be an immediate.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_shuffle_pi16 (__m64 __A, int const __N)
-{
-  unsigned long element_selector_10 = __N & 0x03;
-  unsigned long element_selector_32 = (__N >> 2) & 0x03;
-  unsigned long element_selector_54 = (__N >> 4) & 0x03;
-  unsigned long element_selector_76 = (__N >> 6) & 0x03;
-  static const unsigned short permute_selectors[4] =
-    {
-#ifdef __LITTLE_ENDIAN__
-	      0x0908, 0x0B0A, 0x0D0C, 0x0F0E
-#else
-	      0x0607, 0x0405, 0x0203, 0x0001
-#endif
-    };
-  __m64_union t;
-  __vector unsigned long long a, p, r;
-
-#ifdef __LITTLE_ENDIAN__
-  t.as_short[0] = permute_selectors[element_selector_10];
-  t.as_short[1] = permute_selectors[element_selector_32];
-  t.as_short[2] = permute_selectors[element_selector_54];
-  t.as_short[3] = permute_selectors[element_selector_76];
-#else
-  t.as_short[3] = permute_selectors[element_selector_10];
-  t.as_short[2] = permute_selectors[element_selector_32];
-  t.as_short[1] = permute_selectors[element_selector_54];
-  t.as_short[0] = permute_selectors[element_selector_76];
-#endif
-  p = vec_splats (t.as_m64);
-  a = vec_splats (__A);
-  r = vec_perm (a, a, (__vector unsigned char)p);
-  return (__m64) ((__vector long long) r)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pshufw (__m64 __A, int const __N)
-{
-  return _mm_shuffle_pi16 (__A, __N);
-}
-
-/* Conditionally store byte elements of A into P.  The high bit of each
-   byte in the selector N determines whether the corresponding byte from
-   A is stored.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_maskmove_si64 (__m64 __A, __m64 __N, char *__P)
-{
-  __m64 hibit = 0x8080808080808080UL;
-  __m64 mask, tmp;
-  __m64 *p = (__m64*)__P;
-
-  tmp = *p;
-  mask = _mm_cmpeq_pi8 ((__N & hibit), hibit);
-  tmp = (tmp & (~mask)) | (__A & mask);
-  *p = tmp;
-}
-
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_maskmovq (__m64 __A, __m64 __N, char *__P)
-{
-  _mm_maskmove_si64 (__A, __N, __P);
-}
-
-/* Compute the rounded averages of the unsigned 8-bit values in A and B.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_avg_pu8 (__m64 __A, __m64 __B)
-{
-  __vector unsigned char a, b, c;
-
-  a = (__vector unsigned char)vec_splats (__A);
-  b = (__vector unsigned char)vec_splats (__B);
-  c = vec_avg (a, b);
-  return (__m64) ((__vector long long) c)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pavgb (__m64 __A, __m64 __B)
-{
-  return _mm_avg_pu8 (__A, __B);
-}
-
-/* Compute the rounded averages of the unsigned 16-bit values in A and B.  */
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_avg_pu16 (__m64 __A, __m64 __B)
-{
-  __vector unsigned short a, b, c;
-
-  a = (__vector unsigned short)vec_splats (__A);
-  b = (__vector unsigned short)vec_splats (__B);
-  c = vec_avg (a, b);
-  return (__m64) ((__vector long long) c)[0];
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_pavgw (__m64 __A, __m64 __B)
-{
-  return _mm_avg_pu16 (__A, __B);
-}
-
-/* Compute the sum of the absolute differences of the unsigned 8-bit
-   values in A and B.  Return the value in the lower 16-bit word; the
-   upper words are cleared.  */
-extern __inline    __m64    __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sad_pu8 (__m64  __A, __m64  __B)
-{
-  __vector unsigned char a, b;
-  __vector unsigned char vmin, vmax, vabsdiff;
-  __vector signed int vsum;
-  const __vector unsigned int zero =
-    { 0, 0, 0, 0 };
-  __m64_union result = {0};
-
-  a = (__vector unsigned char) (__vector unsigned long long) { 0UL, __A };
-  b = (__vector unsigned char) (__vector unsigned long long) { 0UL, __B };
-  vmin = vec_min (a, b);
-  vmax = vec_max (a, b);
-  vabsdiff = vec_sub (vmax, vmin);
-  /* Sum four groups of bytes into integers.  */
-  vsum = (__vector signed int) vec_sum4s (vabsdiff, zero);
-  /* Sum across four integers with integer result.  */
-  vsum = vec_sums (vsum, (__vector signed int) zero);
-  /* The sum is in the right most 32-bits of the vector result.
-     Transfer to a GPR and truncate to 16 bits.  */
-  result.as_short[0] = vsum[3];
-  return result.as_m64;
-}
-
-extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_m_psadbw (__m64 __A, __m64 __B)
-{
-  return _mm_sad_pu8 (__A, __B);
-}
-
-/* Stores the data in A to the address P without polluting the caches.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_pi (__m64 *__P, __m64 __A)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "	dcbtstt	0,%0"
-    :
-    : "b" (__P)
-    : "memory"
-  );
-  *__P = __A;
-}
-
-/* Likewise.  The address must be 16-byte aligned.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_stream_ps (float *__P, __m128 __A)
-{
-  /* Use the data cache block touch for store transient.  */
-  __asm__ (
-    "	dcbtstt	0,%0"
-    :
-    : "b" (__P)
-    : "memory"
-  );
-  _mm_store_ps (__P, __A);
-}
-
-/* Guarantees that every preceding store is globally visible before
-   any subsequent store.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_sfence (void)
-{
-  /* Generate a light weight sync.  */
-  __atomic_thread_fence (__ATOMIC_RELEASE);
-}
-
-/* The execution of the next instruction is delayed by an implementation
-   specific amount of time.  The instruction does not modify the
-   architectural state.  This is after the pop_options pragma because
-   it does not require SSE support in the processor--the encoding is a
-   nop on processors that do not support it.  */
-extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-_mm_pause (void)
-{
-  /* There is no exact match with this construct, but the following is
-     close to the desired effect.  */
-#if _ARCH_PWR8
-  /* On power8 and later processors we can depend on Program Priority
-     (PRI) and associated "very low" PPI setting.  Since we don't know
-     what PPI this thread is running at we: 1) save the current PRI
-     from the PPR SPR into a local GRP, 2) set the PRI to "very low*
-     via the special or 31,31,31 encoding. 3) issue an "isync" to
-     insure the PRI change takes effect before we execute any more
-     instructions.
-     Now we can execute a lwsync (release barrier) while we execute
-     this thread at "very low" PRI.  Finally we restore the original
-     PRI and continue execution.  */
-  unsigned long __PPR;
-
-  __asm__ volatile (
-    "	mfppr	%0;"
-    "   or 31,31,31;"
-    "   isync;"
-    "   lwsync;"
-    "   isync;"
-    "   mtppr	%0;"
-    : "=r" (__PPR)
-    :
-    : "memory"
-  );
-#else
-  /* For older processor where we may not even have Program Priority
-     controls we can only depend on Heavy Weight Sync.  */
-  __atomic_thread_fence (__ATOMIC_SEQ_CST);
-#endif
-}
-
-/* Transpose the 4x4 matrix composed of row[0-3].  */
-#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3)			\
-do {									\
-  __v4sf __r0 = (row0), __r1 = (row1), __r2 = (row2), __r3 = (row3);	\
-  __v4sf __t0 = vec_vmrghw (__r0, __r1);			\
-  __v4sf __t1 = vec_vmrghw (__r2, __r3);			\
-  __v4sf __t2 = vec_vmrglw (__r0, __r1);			\
-  __v4sf __t3 = vec_vmrglw (__r2, __r3);			\
-  (row0) = (__v4sf)vec_mergeh ((__vector long long)__t0, 	\
-			       (__vector long long)__t1);	\
-  (row1) = (__v4sf)vec_mergel ((__vector long long)__t0,	\
-			       (__vector long long)__t1);	\
-  (row2) = (__v4sf)vec_mergeh ((__vector long long)__t2,	\
-			       (__vector long long)__t3);	\
-  (row3) = (__v4sf)vec_mergel ((__vector long long)__t2,	\
-			       (__vector long long)__t3);	\
-} while (0)
-
-/* For backward source compatibility.  */
-//# include <emmintrin.h>
-
-#endif /* _XMMINTRIN_H_INCLUDED */
diff --git a/linux-x86/lib64/clang/9.0.7/include/vecintrin.h b/linux-x86/lib64/clang/9.0.7/include/vecintrin.h
deleted file mode 100644
index 115d893..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/vecintrin.h
+++ /dev/null
@@ -1,10456 +0,0 @@
-/*===---- vecintrin.h - Vector intrinsics ----------------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#if defined(__s390x__) && defined(__VEC__)
-
-#define __ATTRS_ai __attribute__((__always_inline__))
-#define __ATTRS_o __attribute__((__overloadable__))
-#define __ATTRS_o_ai __attribute__((__overloadable__, __always_inline__))
-
-#define __constant(PARM) \
-  __attribute__((__enable_if__ ((PARM) == (PARM), \
-     "argument must be a constant integer")))
-#define __constant_range(PARM, LOW, HIGH) \
-  __attribute__((__enable_if__ ((PARM) >= (LOW) && (PARM) <= (HIGH), \
-     "argument must be a constant integer from " #LOW " to " #HIGH)))
-#define __constant_pow2_range(PARM, LOW, HIGH) \
-  __attribute__((__enable_if__ ((PARM) >= (LOW) && (PARM) <= (HIGH) && \
-                                ((PARM) & ((PARM) - 1)) == 0, \
-     "argument must be a constant power of 2 from " #LOW " to " #HIGH)))
-
-/*-- __lcbb -----------------------------------------------------------------*/
-
-extern __ATTRS_o unsigned int
-__lcbb(const void *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-#define __lcbb(X, Y) ((__typeof__((__lcbb)((X), (Y)))) \
-  __builtin_s390_lcbb((X), __builtin_constant_p((Y))? \
-                           ((Y) == 64 ? 0 : \
-                            (Y) == 128 ? 1 : \
-                            (Y) == 256 ? 2 : \
-                            (Y) == 512 ? 3 : \
-                            (Y) == 1024 ? 4 : \
-                            (Y) == 2048 ? 5 : \
-                            (Y) == 4096 ? 6 : 0) : 0))
-
-/*-- vec_extract ------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai signed char
-vec_extract(vector signed char __vec, int __index) {
-  return __vec[__index & 15];
-}
-
-static inline __ATTRS_o_ai unsigned char
-vec_extract(vector bool char __vec, int __index) {
-  return __vec[__index & 15];
-}
-
-static inline __ATTRS_o_ai unsigned char
-vec_extract(vector unsigned char __vec, int __index) {
-  return __vec[__index & 15];
-}
-
-static inline __ATTRS_o_ai signed short
-vec_extract(vector signed short __vec, int __index) {
-  return __vec[__index & 7];
-}
-
-static inline __ATTRS_o_ai unsigned short
-vec_extract(vector bool short __vec, int __index) {
-  return __vec[__index & 7];
-}
-
-static inline __ATTRS_o_ai unsigned short
-vec_extract(vector unsigned short __vec, int __index) {
-  return __vec[__index & 7];
-}
-
-static inline __ATTRS_o_ai signed int
-vec_extract(vector signed int __vec, int __index) {
-  return __vec[__index & 3];
-}
-
-static inline __ATTRS_o_ai unsigned int
-vec_extract(vector bool int __vec, int __index) {
-  return __vec[__index & 3];
-}
-
-static inline __ATTRS_o_ai unsigned int
-vec_extract(vector unsigned int __vec, int __index) {
-  return __vec[__index & 3];
-}
-
-static inline __ATTRS_o_ai signed long long
-vec_extract(vector signed long long __vec, int __index) {
-  return __vec[__index & 1];
-}
-
-static inline __ATTRS_o_ai unsigned long long
-vec_extract(vector bool long long __vec, int __index) {
-  return __vec[__index & 1];
-}
-
-static inline __ATTRS_o_ai unsigned long long
-vec_extract(vector unsigned long long __vec, int __index) {
-  return __vec[__index & 1];
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai float
-vec_extract(vector float __vec, int __index) {
-  return __vec[__index & 3];
-}
-#endif
-
-static inline __ATTRS_o_ai double
-vec_extract(vector double __vec, int __index) {
-  return __vec[__index & 1];
-}
-
-/*-- vec_insert -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_insert(signed char __scalar, vector signed char __vec, int __index) {
-  __vec[__index & 15] = __scalar;
-  return __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_insert(unsigned char __scalar, vector bool char __vec, int __index) {
-  vector unsigned char __newvec = (vector unsigned char)__vec;
-  __newvec[__index & 15] = (unsigned char)__scalar;
-  return __newvec;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_insert(unsigned char __scalar, vector unsigned char __vec, int __index) {
-  __vec[__index & 15] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_insert(signed short __scalar, vector signed short __vec, int __index) {
-  __vec[__index & 7] = __scalar;
-  return __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_insert(unsigned short __scalar, vector bool short __vec, int __index) {
-  vector unsigned short __newvec = (vector unsigned short)__vec;
-  __newvec[__index & 7] = (unsigned short)__scalar;
-  return __newvec;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_insert(unsigned short __scalar, vector unsigned short __vec, int __index) {
-  __vec[__index & 7] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_insert(signed int __scalar, vector signed int __vec, int __index) {
-  __vec[__index & 3] = __scalar;
-  return __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_insert(unsigned int __scalar, vector bool int __vec, int __index) {
-  vector unsigned int __newvec = (vector unsigned int)__vec;
-  __newvec[__index & 3] = __scalar;
-  return __newvec;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_insert(unsigned int __scalar, vector unsigned int __vec, int __index) {
-  __vec[__index & 3] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_insert(signed long long __scalar, vector signed long long __vec,
-           int __index) {
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_insert(unsigned long long __scalar, vector bool long long __vec,
-           int __index) {
-  vector unsigned long long __newvec = (vector unsigned long long)__vec;
-  __newvec[__index & 1] = __scalar;
-  return __newvec;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_insert(unsigned long long __scalar, vector unsigned long long __vec,
-           int __index) {
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_insert(float __scalar, vector float __vec, int __index) {
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_insert(double __scalar, vector double __vec, int __index) {
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-/*-- vec_promote ------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_promote(signed char __scalar, int __index) {
-  const vector signed char __zero = (vector signed char)0;
-  vector signed char __vec = __builtin_shufflevector(__zero, __zero,
-    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1);
-  __vec[__index & 15] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_promote(unsigned char __scalar, int __index) {
-  const vector unsigned char __zero = (vector unsigned char)0;
-  vector unsigned char __vec = __builtin_shufflevector(__zero, __zero,
-    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1);
-  __vec[__index & 15] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_promote(signed short __scalar, int __index) {
-  const vector signed short __zero = (vector signed short)0;
-  vector signed short __vec = __builtin_shufflevector(__zero, __zero,
-                                -1, -1, -1, -1, -1, -1, -1, -1);
-  __vec[__index & 7] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_promote(unsigned short __scalar, int __index) {
-  const vector unsigned short __zero = (vector unsigned short)0;
-  vector unsigned short __vec = __builtin_shufflevector(__zero, __zero,
-                                  -1, -1, -1, -1, -1, -1, -1, -1);
-  __vec[__index & 7] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_promote(signed int __scalar, int __index) {
-  const vector signed int __zero = (vector signed int)0;
-  vector signed int __vec = __builtin_shufflevector(__zero, __zero,
-                                                    -1, -1, -1, -1);
-  __vec[__index & 3] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_promote(unsigned int __scalar, int __index) {
-  const vector unsigned int __zero = (vector unsigned int)0;
-  vector unsigned int __vec = __builtin_shufflevector(__zero, __zero,
-                                                      -1, -1, -1, -1);
-  __vec[__index & 3] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_promote(signed long long __scalar, int __index) {
-  const vector signed long long __zero = (vector signed long long)0;
-  vector signed long long __vec = __builtin_shufflevector(__zero, __zero,
-                                                          -1, -1);
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_promote(unsigned long long __scalar, int __index) {
-  const vector unsigned long long __zero = (vector unsigned long long)0;
-  vector unsigned long long __vec = __builtin_shufflevector(__zero, __zero,
-                                                            -1, -1);
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_promote(float __scalar, int __index) {
-  const vector float __zero = (vector float)0;
-  vector float __vec = __builtin_shufflevector(__zero, __zero, -1, -1, -1, -1);
-  __vec[__index & 3] = __scalar;
-  return __vec;
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_promote(double __scalar, int __index) {
-  const vector double __zero = (vector double)0;
-  vector double __vec = __builtin_shufflevector(__zero, __zero, -1, -1);
-  __vec[__index & 1] = __scalar;
-  return __vec;
-}
-
-/*-- vec_insert_and_zero ----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_insert_and_zero(const signed char *__ptr) {
-  vector signed char __vec = (vector signed char)0;
-  __vec[7] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_insert_and_zero(const unsigned char *__ptr) {
-  vector unsigned char __vec = (vector unsigned char)0;
-  __vec[7] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_insert_and_zero(const signed short *__ptr) {
-  vector signed short __vec = (vector signed short)0;
-  __vec[3] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_insert_and_zero(const unsigned short *__ptr) {
-  vector unsigned short __vec = (vector unsigned short)0;
-  __vec[3] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_insert_and_zero(const signed int *__ptr) {
-  vector signed int __vec = (vector signed int)0;
-  __vec[1] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_insert_and_zero(const unsigned int *__ptr) {
-  vector unsigned int __vec = (vector unsigned int)0;
-  __vec[1] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_insert_and_zero(const signed long long *__ptr) {
-  vector signed long long __vec = (vector signed long long)0;
-  __vec[0] = *__ptr;
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_insert_and_zero(const unsigned long long *__ptr) {
-  vector unsigned long long __vec = (vector unsigned long long)0;
-  __vec[0] = *__ptr;
-  return __vec;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_insert_and_zero(const float *__ptr) {
-  vector float __vec = (vector float)0;
-  __vec[1] = *__ptr;
-  return __vec;
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_insert_and_zero(const double *__ptr) {
-  vector double __vec = (vector double)0;
-  __vec[0] = *__ptr;
-  return __vec;
-}
-
-/*-- vec_perm ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_perm(vector signed char __a, vector signed char __b,
-         vector unsigned char __c) {
-  return (vector signed char)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_perm(vector unsigned char __a, vector unsigned char __b,
-         vector unsigned char __c) {
-  return (vector unsigned char)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_perm(vector bool char __a, vector bool char __b,
-         vector unsigned char __c) {
-  return (vector bool char)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_perm(vector signed short __a, vector signed short __b,
-         vector unsigned char __c) {
-  return (vector signed short)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_perm(vector unsigned short __a, vector unsigned short __b,
-         vector unsigned char __c) {
-  return (vector unsigned short)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_perm(vector bool short __a, vector bool short __b,
-         vector unsigned char __c) {
-  return (vector bool short)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_perm(vector signed int __a, vector signed int __b,
-         vector unsigned char __c) {
-  return (vector signed int)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_perm(vector unsigned int __a, vector unsigned int __b,
-         vector unsigned char __c) {
-  return (vector unsigned int)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_perm(vector bool int __a, vector bool int __b,
-         vector unsigned char __c) {
-  return (vector bool int)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_perm(vector signed long long __a, vector signed long long __b,
-         vector unsigned char __c) {
-  return (vector signed long long)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_perm(vector unsigned long long __a, vector unsigned long long __b,
-         vector unsigned char __c) {
-  return (vector unsigned long long)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_perm(vector bool long long __a, vector bool long long __b,
-         vector unsigned char __c) {
-  return (vector bool long long)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_perm(vector float __a, vector float __b,
-         vector unsigned char __c) {
-  return (vector float)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_perm(vector double __a, vector double __b,
-         vector unsigned char __c) {
-  return (vector double)__builtin_s390_vperm(
-           (vector unsigned char)__a, (vector unsigned char)__b, __c);
-}
-
-/*-- vec_permi --------------------------------------------------------------*/
-
-// This prototype is deprecated.
-extern __ATTRS_o vector signed long long
-vec_permi(vector signed long long __a, vector signed long long __b, int __c)
-  __constant_range(__c, 0, 3);
-
-// This prototype is deprecated.
-extern __ATTRS_o vector unsigned long long
-vec_permi(vector unsigned long long __a, vector unsigned long long __b, int __c)
-  __constant_range(__c, 0, 3);
-
-// This prototype is deprecated.
-extern __ATTRS_o vector bool long long
-vec_permi(vector bool long long __a, vector bool long long __b, int __c)
-  __constant_range(__c, 0, 3);
-
-// This prototype is deprecated.
-extern __ATTRS_o vector double
-vec_permi(vector double __a, vector double __b, int __c)
-  __constant_range(__c, 0, 3);
-
-#define vec_permi(X, Y, Z) ((__typeof__((vec_permi)((X), (Y), (Z)))) \
-  __builtin_s390_vpdi((vector unsigned long long)(X), \
-                      (vector unsigned long long)(Y), \
-                      (((Z) & 2) << 1) | ((Z) & 1)))
-
-/*-- vec_bperm_u128 ---------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_ai vector unsigned long long
-vec_bperm_u128(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vbperm(__a, __b);
-}
-#endif
-
-/*-- vec_sel ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_sel(vector signed char __a, vector signed char __b,
-        vector unsigned char __c) {
-  return ((vector signed char)__c & __b) | (~(vector signed char)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_sel(vector signed char __a, vector signed char __b, vector bool char __c) {
-  return ((vector signed char)__c & __b) | (~(vector signed char)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_sel(vector bool char __a, vector bool char __b, vector unsigned char __c) {
-  return ((vector bool char)__c & __b) | (~(vector bool char)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_sel(vector bool char __a, vector bool char __b, vector bool char __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sel(vector unsigned char __a, vector unsigned char __b,
-        vector unsigned char __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sel(vector unsigned char __a, vector unsigned char __b,
-        vector bool char __c) {
-  return ((vector unsigned char)__c & __b) | (~(vector unsigned char)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_sel(vector signed short __a, vector signed short __b,
-        vector unsigned short __c) {
-  return ((vector signed short)__c & __b) | (~(vector signed short)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_sel(vector signed short __a, vector signed short __b,
-        vector bool short __c) {
-  return ((vector signed short)__c & __b) | (~(vector signed short)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_sel(vector bool short __a, vector bool short __b,
-        vector unsigned short __c) {
-  return ((vector bool short)__c & __b) | (~(vector bool short)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_sel(vector bool short __a, vector bool short __b, vector bool short __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_sel(vector unsigned short __a, vector unsigned short __b,
-        vector unsigned short __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_sel(vector unsigned short __a, vector unsigned short __b,
-        vector bool short __c) {
-  return (((vector unsigned short)__c & __b) |
-          (~(vector unsigned short)__c & __a));
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_sel(vector signed int __a, vector signed int __b,
-        vector unsigned int __c) {
-  return ((vector signed int)__c & __b) | (~(vector signed int)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_sel(vector signed int __a, vector signed int __b, vector bool int __c) {
-  return ((vector signed int)__c & __b) | (~(vector signed int)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_sel(vector bool int __a, vector bool int __b, vector unsigned int __c) {
-  return ((vector bool int)__c & __b) | (~(vector bool int)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_sel(vector bool int __a, vector bool int __b, vector bool int __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sel(vector unsigned int __a, vector unsigned int __b,
-        vector unsigned int __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sel(vector unsigned int __a, vector unsigned int __b, vector bool int __c) {
-  return ((vector unsigned int)__c & __b) | (~(vector unsigned int)__c & __a);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_sel(vector signed long long __a, vector signed long long __b,
-        vector unsigned long long __c) {
-  return (((vector signed long long)__c & __b) |
-          (~(vector signed long long)__c & __a));
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_sel(vector signed long long __a, vector signed long long __b,
-        vector bool long long __c) {
-  return (((vector signed long long)__c & __b) |
-          (~(vector signed long long)__c & __a));
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_sel(vector bool long long __a, vector bool long long __b,
-        vector unsigned long long __c) {
-  return (((vector bool long long)__c & __b) |
-          (~(vector bool long long)__c & __a));
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_sel(vector bool long long __a, vector bool long long __b,
-        vector bool long long __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sel(vector unsigned long long __a, vector unsigned long long __b,
-        vector unsigned long long __c) {
-  return (__c & __b) | (~__c & __a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sel(vector unsigned long long __a, vector unsigned long long __b,
-        vector bool long long __c) {
-  return (((vector unsigned long long)__c & __b) |
-          (~(vector unsigned long long)__c & __a));
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_sel(vector float __a, vector float __b, vector unsigned int __c) {
-  return (vector float)((__c & (vector unsigned int)__b) |
-                        (~__c & (vector unsigned int)__a));
-}
-
-static inline __ATTRS_o_ai vector float
-vec_sel(vector float __a, vector float __b, vector bool int __c) {
-  vector unsigned int __ac = (vector unsigned int)__a;
-  vector unsigned int __bc = (vector unsigned int)__b;
-  vector unsigned int __cc = (vector unsigned int)__c;
-  return (vector float)((__cc & __bc) | (~__cc & __ac));
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_sel(vector double __a, vector double __b, vector unsigned long long __c) {
-  return (vector double)((__c & (vector unsigned long long)__b) |
-                         (~__c & (vector unsigned long long)__a));
-}
-
-static inline __ATTRS_o_ai vector double
-vec_sel(vector double __a, vector double __b, vector bool long long __c) {
-  vector unsigned long long __ac = (vector unsigned long long)__a;
-  vector unsigned long long __bc = (vector unsigned long long)__b;
-  vector unsigned long long __cc = (vector unsigned long long)__c;
-  return (vector double)((__cc & __bc) | (~__cc & __ac));
-}
-
-/*-- vec_gather_element -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed int
-vec_gather_element(vector signed int __vec, vector unsigned int __offset,
-                   const signed int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  __vec[__index] = *(const signed int *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_gather_element(vector bool int __vec, vector unsigned int __offset,
-                   const unsigned int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  __vec[__index] = *(const unsigned int *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_gather_element(vector unsigned int __vec, vector unsigned int __offset,
-                   const unsigned int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  __vec[__index] = *(const unsigned int *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_gather_element(vector signed long long __vec,
-                   vector unsigned long long __offset,
-                   const signed long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  __vec[__index] = *(const signed long long *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_gather_element(vector bool long long __vec,
-                   vector unsigned long long __offset,
-                   const unsigned long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  __vec[__index] = *(const unsigned long long *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_gather_element(vector unsigned long long __vec,
-                   vector unsigned long long __offset,
-                   const unsigned long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  __vec[__index] = *(const unsigned long long *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_gather_element(vector float __vec, vector unsigned int __offset,
-                   const float *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  __vec[__index] = *(const float *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_gather_element(vector double __vec, vector unsigned long long __offset,
-                   const double *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  __vec[__index] = *(const double *)(
-    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
-  return __vec;
-}
-
-/*-- vec_scatter_element ----------------------------------------------------*/
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector signed int __vec, vector unsigned int __offset,
-                    signed int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  *(signed int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector bool int __vec, vector unsigned int __offset,
-                    unsigned int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  *(unsigned int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector unsigned int __vec, vector unsigned int __offset,
-                    unsigned int *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  *(unsigned int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector signed long long __vec,
-                    vector unsigned long long __offset,
-                    signed long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  *(signed long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector bool long long __vec,
-                    vector unsigned long long __offset,
-                    unsigned long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  *(unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector unsigned long long __vec,
-                    vector unsigned long long __offset,
-                    unsigned long long *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  *(unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector float __vec, vector unsigned int __offset,
-                    float *__ptr, int __index)
-  __constant_range(__index, 0, 3) {
-  *(float *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-#endif
-
-static inline __ATTRS_o_ai void
-vec_scatter_element(vector double __vec, vector unsigned long long __offset,
-                    double *__ptr, int __index)
-  __constant_range(__index, 0, 1) {
-  *(double *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
-    __vec[__index];
-}
-
-/*-- vec_xl -----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_xl(long __offset, const signed char *__ptr) {
-  return *(const vector signed char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_xl(long __offset, const unsigned char *__ptr) {
-  return *(const vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_xl(long __offset, const signed short *__ptr) {
-  return *(const vector signed short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_xl(long __offset, const unsigned short *__ptr) {
-  return *(const vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_xl(long __offset, const signed int *__ptr) {
-  return *(const vector signed int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_xl(long __offset, const unsigned int *__ptr) {
-  return *(const vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_xl(long __offset, const signed long long *__ptr) {
-  return *(const vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_xl(long __offset, const unsigned long long *__ptr) {
-  return *(const vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_xl(long __offset, const float *__ptr) {
-  return *(const vector float *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_xl(long __offset, const double *__ptr) {
-  return *(const vector double *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-/*-- vec_xld2 ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_xld2(long __offset, const signed char *__ptr) {
-  return *(const vector signed char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_xld2(long __offset, const unsigned char *__ptr) {
-  return *(const vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_xld2(long __offset, const signed short *__ptr) {
-  return *(const vector signed short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_xld2(long __offset, const unsigned short *__ptr) {
-  return *(const vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_xld2(long __offset, const signed int *__ptr) {
-  return *(const vector signed int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_xld2(long __offset, const unsigned int *__ptr) {
-  return *(const vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_xld2(long __offset, const signed long long *__ptr) {
-  return *(const vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_xld2(long __offset, const unsigned long long *__ptr) {
-  return *(const vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_xld2(long __offset, const double *__ptr) {
-  return *(const vector double *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-/*-- vec_xlw4 ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_xlw4(long __offset, const signed char *__ptr) {
-  return *(const vector signed char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_xlw4(long __offset, const unsigned char *__ptr) {
-  return *(const vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_xlw4(long __offset, const signed short *__ptr) {
-  return *(const vector signed short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_xlw4(long __offset, const unsigned short *__ptr) {
-  return *(const vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_xlw4(long __offset, const signed int *__ptr) {
-  return *(const vector signed int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_xlw4(long __offset, const unsigned int *__ptr) {
-  return *(const vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset);
-}
-
-/*-- vec_xst ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai void
-vec_xst(vector signed char __vec, long __offset, signed char *__ptr) {
-  *(vector signed char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector signed short __vec, long __offset, signed short *__ptr) {
-  *(vector signed short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector signed int __vec, long __offset, signed int *__ptr) {
-  *(vector signed int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector signed long long __vec, long __offset,
-          signed long long *__ptr) {
-  *(vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-static inline __ATTRS_o_ai void
-vec_xst(vector unsigned long long __vec, long __offset,
-          unsigned long long *__ptr) {
-  *(vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset) =
-    __vec;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai void
-vec_xst(vector float __vec, long __offset, float *__ptr) {
-  *(vector float *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-#endif
-
-static inline __ATTRS_o_ai void
-vec_xst(vector double __vec, long __offset, double *__ptr) {
-  *(vector double *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-/*-- vec_xstd2 --------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector signed char __vec, long __offset, signed char *__ptr) {
-  *(vector signed char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector signed short __vec, long __offset, signed short *__ptr) {
-  *(vector signed short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector signed int __vec, long __offset, signed int *__ptr) {
-  *(vector signed int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector signed long long __vec, long __offset,
-          signed long long *__ptr) {
-  *(vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector unsigned long long __vec, long __offset,
-          unsigned long long *__ptr) {
-  *(vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset) =
-    __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstd2(vector double __vec, long __offset, double *__ptr) {
-  *(vector double *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-/*-- vec_xstw4 --------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector signed char __vec, long __offset, signed char *__ptr) {
-  *(vector signed char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector unsigned char __vec, long __offset, unsigned char *__ptr) {
-  *(vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector signed short __vec, long __offset, signed short *__ptr) {
-  *(vector signed short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector unsigned short __vec, long __offset, unsigned short *__ptr) {
-  *(vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector signed int __vec, long __offset, signed int *__ptr) {
-  *(vector signed int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai void
-vec_xstw4(vector unsigned int __vec, long __offset, unsigned int *__ptr) {
-  *(vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
-}
-
-/*-- vec_load_bndry ---------------------------------------------------------*/
-
-extern __ATTRS_o vector signed char
-vec_load_bndry(const signed char *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector unsigned char
-vec_load_bndry(const unsigned char *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector signed short
-vec_load_bndry(const signed short *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector unsigned short
-vec_load_bndry(const unsigned short *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector signed int
-vec_load_bndry(const signed int *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector unsigned int
-vec_load_bndry(const unsigned int *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector signed long long
-vec_load_bndry(const signed long long *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-extern __ATTRS_o vector unsigned long long
-vec_load_bndry(const unsigned long long *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-#if __ARCH__ >= 12
-extern __ATTRS_o vector float
-vec_load_bndry(const float *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-#endif
-
-extern __ATTRS_o vector double
-vec_load_bndry(const double *__ptr, unsigned short __len)
-  __constant_pow2_range(__len, 64, 4096);
-
-#define vec_load_bndry(X, Y) ((__typeof__((vec_load_bndry)((X), (Y)))) \
-  __builtin_s390_vlbb((X), ((Y) == 64 ? 0 : \
-                            (Y) == 128 ? 1 : \
-                            (Y) == 256 ? 2 : \
-                            (Y) == 512 ? 3 : \
-                            (Y) == 1024 ? 4 : \
-                            (Y) == 2048 ? 5 : \
-                            (Y) == 4096 ? 6 : -1)))
-
-/*-- vec_load_len -----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_load_len(const signed char *__ptr, unsigned int __len) {
-  return (vector signed char)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_load_len(const unsigned char *__ptr, unsigned int __len) {
-  return (vector unsigned char)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_load_len(const signed short *__ptr, unsigned int __len) {
-  return (vector signed short)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_load_len(const unsigned short *__ptr, unsigned int __len) {
-  return (vector unsigned short)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_load_len(const signed int *__ptr, unsigned int __len) {
-  return (vector signed int)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_load_len(const unsigned int *__ptr, unsigned int __len) {
-  return (vector unsigned int)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_load_len(const signed long long *__ptr, unsigned int __len) {
-  return (vector signed long long)__builtin_s390_vll(__len, __ptr);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_load_len(const unsigned long long *__ptr, unsigned int __len) {
-  return (vector unsigned long long)__builtin_s390_vll(__len, __ptr);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_load_len(const float *__ptr, unsigned int __len) {
-  return (vector float)__builtin_s390_vll(__len, __ptr);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_load_len(const double *__ptr, unsigned int __len) {
-  return (vector double)__builtin_s390_vll(__len, __ptr);
-}
-
-/*-- vec_load_len_r ---------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_ai vector unsigned char
-vec_load_len_r(const unsigned char *__ptr, unsigned int __len) {
-  return (vector unsigned char)__builtin_s390_vlrl(__len, __ptr);
-}
-#endif
-
-/*-- vec_store_len ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector signed char __vec, signed char *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector unsigned char __vec, unsigned char *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector signed short __vec, signed short *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector unsigned short __vec, unsigned short *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector signed int __vec, signed int *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector unsigned int __vec, unsigned int *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector signed long long __vec, signed long long *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector unsigned long long __vec, unsigned long long *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai void
-vec_store_len(vector float __vec, float *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-#endif
-
-static inline __ATTRS_o_ai void
-vec_store_len(vector double __vec, double *__ptr,
-              unsigned int __len) {
-  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
-}
-
-/*-- vec_store_len_r --------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_ai void
-vec_store_len_r(vector unsigned char __vec, unsigned char *__ptr,
-                unsigned int __len) {
-  __builtin_s390_vstrl((vector signed char)__vec, __len, __ptr);
-}
-#endif
-
-/*-- vec_load_pair ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed long long
-vec_load_pair(signed long long __a, signed long long __b) {
-  return (vector signed long long)(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_load_pair(unsigned long long __a, unsigned long long __b) {
-  return (vector unsigned long long)(__a, __b);
-}
-
-/*-- vec_genmask ------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_genmask(unsigned short __mask)
-  __constant(__mask) {
-  return (vector unsigned char)(
-    __mask & 0x8000 ? 0xff : 0,
-    __mask & 0x4000 ? 0xff : 0,
-    __mask & 0x2000 ? 0xff : 0,
-    __mask & 0x1000 ? 0xff : 0,
-    __mask & 0x0800 ? 0xff : 0,
-    __mask & 0x0400 ? 0xff : 0,
-    __mask & 0x0200 ? 0xff : 0,
-    __mask & 0x0100 ? 0xff : 0,
-    __mask & 0x0080 ? 0xff : 0,
-    __mask & 0x0040 ? 0xff : 0,
-    __mask & 0x0020 ? 0xff : 0,
-    __mask & 0x0010 ? 0xff : 0,
-    __mask & 0x0008 ? 0xff : 0,
-    __mask & 0x0004 ? 0xff : 0,
-    __mask & 0x0002 ? 0xff : 0,
-    __mask & 0x0001 ? 0xff : 0);
-}
-
-/*-- vec_genmasks_* ---------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_genmasks_8(unsigned char __first, unsigned char __last)
-  __constant(__first) __constant(__last) {
-  unsigned char __bit1 = __first & 7;
-  unsigned char __bit2 = __last & 7;
-  unsigned char __mask1 = (unsigned char)(1U << (7 - __bit1) << 1) - 1;
-  unsigned char __mask2 = (unsigned char)(1U << (7 - __bit2)) - 1;
-  unsigned char __value = (__bit1 <= __bit2 ?
-                           __mask1 & ~__mask2 :
-                           __mask1 | ~__mask2);
-  return (vector unsigned char)__value;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_genmasks_16(unsigned char __first, unsigned char __last)
-  __constant(__first) __constant(__last) {
-  unsigned char __bit1 = __first & 15;
-  unsigned char __bit2 = __last & 15;
-  unsigned short __mask1 = (unsigned short)(1U << (15 - __bit1) << 1) - 1;
-  unsigned short __mask2 = (unsigned short)(1U << (15 - __bit2)) - 1;
-  unsigned short __value = (__bit1 <= __bit2 ?
-                            __mask1 & ~__mask2 :
-                            __mask1 | ~__mask2);
-  return (vector unsigned short)__value;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_genmasks_32(unsigned char __first, unsigned char __last)
-  __constant(__first) __constant(__last) {
-  unsigned char __bit1 = __first & 31;
-  unsigned char __bit2 = __last & 31;
-  unsigned int __mask1 = (1U << (31 - __bit1) << 1) - 1;
-  unsigned int __mask2 = (1U << (31 - __bit2)) - 1;
-  unsigned int __value = (__bit1 <= __bit2 ?
-                          __mask1 & ~__mask2 :
-                          __mask1 | ~__mask2);
-  return (vector unsigned int)__value;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_genmasks_64(unsigned char __first, unsigned char __last)
-  __constant(__first) __constant(__last) {
-  unsigned char __bit1 = __first & 63;
-  unsigned char __bit2 = __last & 63;
-  unsigned long long __mask1 = (1ULL << (63 - __bit1) << 1) - 1;
-  unsigned long long __mask2 = (1ULL << (63 - __bit2)) - 1;
-  unsigned long long __value = (__bit1 <= __bit2 ?
-                                __mask1 & ~__mask2 :
-                                __mask1 | ~__mask2);
-  return (vector unsigned long long)__value;
-}
-
-/*-- vec_splat --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_splat(vector signed char __vec, int __index)
-  __constant_range(__index, 0, 15) {
-  return (vector signed char)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_splat(vector bool char __vec, int __index)
-  __constant_range(__index, 0, 15) {
-  return (vector bool char)(vector unsigned char)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_splat(vector unsigned char __vec, int __index)
-  __constant_range(__index, 0, 15) {
-  return (vector unsigned char)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_splat(vector signed short __vec, int __index)
-  __constant_range(__index, 0, 7) {
-  return (vector signed short)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_splat(vector bool short __vec, int __index)
-  __constant_range(__index, 0, 7) {
-  return (vector bool short)(vector unsigned short)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_splat(vector unsigned short __vec, int __index)
-  __constant_range(__index, 0, 7) {
-  return (vector unsigned short)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_splat(vector signed int __vec, int __index)
-  __constant_range(__index, 0, 3) {
-  return (vector signed int)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_splat(vector bool int __vec, int __index)
-  __constant_range(__index, 0, 3) {
-  return (vector bool int)(vector unsigned int)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_splat(vector unsigned int __vec, int __index)
-  __constant_range(__index, 0, 3) {
-  return (vector unsigned int)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_splat(vector signed long long __vec, int __index)
-  __constant_range(__index, 0, 1) {
-  return (vector signed long long)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_splat(vector bool long long __vec, int __index)
-  __constant_range(__index, 0, 1) {
-  return (vector bool long long)(vector unsigned long long)__vec[__index];
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_splat(vector unsigned long long __vec, int __index)
-  __constant_range(__index, 0, 1) {
-  return (vector unsigned long long)__vec[__index];
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_splat(vector float __vec, int __index)
-  __constant_range(__index, 0, 3) {
-  return (vector float)__vec[__index];
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_splat(vector double __vec, int __index)
-  __constant_range(__index, 0, 1) {
-  return (vector double)__vec[__index];
-}
-
-/*-- vec_splat_s* -----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector signed char
-vec_splat_s8(signed char __scalar)
-  __constant(__scalar) {
-  return (vector signed char)__scalar;
-}
-
-static inline __ATTRS_ai vector signed short
-vec_splat_s16(signed short __scalar)
-  __constant(__scalar) {
-  return (vector signed short)__scalar;
-}
-
-static inline __ATTRS_ai vector signed int
-vec_splat_s32(signed short __scalar)
-  __constant(__scalar) {
-  return (vector signed int)(signed int)__scalar;
-}
-
-static inline __ATTRS_ai vector signed long long
-vec_splat_s64(signed short __scalar)
-  __constant(__scalar) {
-  return (vector signed long long)(signed long)__scalar;
-}
-
-/*-- vec_splat_u* -----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_splat_u8(unsigned char __scalar)
-  __constant(__scalar) {
-  return (vector unsigned char)__scalar;
-}
-
-static inline __ATTRS_ai vector unsigned short
-vec_splat_u16(unsigned short __scalar)
-  __constant(__scalar) {
-  return (vector unsigned short)__scalar;
-}
-
-static inline __ATTRS_ai vector unsigned int
-vec_splat_u32(signed short __scalar)
-  __constant(__scalar) {
-  return (vector unsigned int)(signed int)__scalar;
-}
-
-static inline __ATTRS_ai vector unsigned long long
-vec_splat_u64(signed short __scalar)
-  __constant(__scalar) {
-  return (vector unsigned long long)(signed long long)__scalar;
-}
-
-/*-- vec_splats -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_splats(signed char __scalar) {
-  return (vector signed char)__scalar;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_splats(unsigned char __scalar) {
-  return (vector unsigned char)__scalar;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_splats(signed short __scalar) {
-  return (vector signed short)__scalar;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_splats(unsigned short __scalar) {
-  return (vector unsigned short)__scalar;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_splats(signed int __scalar) {
-  return (vector signed int)__scalar;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_splats(unsigned int __scalar) {
-  return (vector unsigned int)__scalar;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_splats(signed long long __scalar) {
-  return (vector signed long long)__scalar;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_splats(unsigned long long __scalar) {
-  return (vector unsigned long long)__scalar;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_splats(float __scalar) {
-  return (vector float)__scalar;
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_splats(double __scalar) {
-  return (vector double)__scalar;
-}
-
-/*-- vec_extend_s64 ---------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed long long
-vec_extend_s64(vector signed char __a) {
-  return (vector signed long long)(__a[7], __a[15]);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_extend_s64(vector signed short __a) {
-  return (vector signed long long)(__a[3], __a[7]);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_extend_s64(vector signed int __a) {
-  return (vector signed long long)(__a[1], __a[3]);
-}
-
-/*-- vec_mergeh -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_mergeh(vector signed char __a, vector signed char __b) {
-  return (vector signed char)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_mergeh(vector bool char __a, vector bool char __b) {
-  return (vector bool char)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_mergeh(vector unsigned char __a, vector unsigned char __b) {
-  return (vector unsigned char)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mergeh(vector signed short __a, vector signed short __b) {
-  return (vector signed short)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_mergeh(vector bool short __a, vector bool short __b) {
-  return (vector bool short)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mergeh(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)(
-    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mergeh(vector signed int __a, vector signed int __b) {
-  return (vector signed int)(__a[0], __b[0], __a[1], __b[1]);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_mergeh(vector bool int __a, vector bool int __b) {
-  return (vector bool int)(__a[0], __b[0], __a[1], __b[1]);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mergeh(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)(__a[0], __b[0], __a[1], __b[1]);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_mergeh(vector signed long long __a, vector signed long long __b) {
-  return (vector signed long long)(__a[0], __b[0]);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_mergeh(vector bool long long __a, vector bool long long __b) {
-  return (vector bool long long)(__a[0], __b[0]);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_mergeh(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector unsigned long long)(__a[0], __b[0]);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_mergeh(vector float __a, vector float __b) {
-  return (vector float)(__a[0], __b[0], __a[1], __b[1]);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_mergeh(vector double __a, vector double __b) {
-  return (vector double)(__a[0], __b[0]);
-}
-
-/*-- vec_mergel -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_mergel(vector signed char __a, vector signed char __b) {
-  return (vector signed char)(
-    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
-    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_mergel(vector bool char __a, vector bool char __b) {
-  return (vector bool char)(
-    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
-    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_mergel(vector unsigned char __a, vector unsigned char __b) {
-  return (vector unsigned char)(
-    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
-    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mergel(vector signed short __a, vector signed short __b) {
-  return (vector signed short)(
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_mergel(vector bool short __a, vector bool short __b) {
-  return (vector bool short)(
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mergel(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)(
-    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mergel(vector signed int __a, vector signed int __b) {
-  return (vector signed int)(__a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_mergel(vector bool int __a, vector bool int __b) {
-  return (vector bool int)(__a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mergel(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)(__a[2], __b[2], __a[3], __b[3]);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_mergel(vector signed long long __a, vector signed long long __b) {
-  return (vector signed long long)(__a[1], __b[1]);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_mergel(vector bool long long __a, vector bool long long __b) {
-  return (vector bool long long)(__a[1], __b[1]);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_mergel(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector unsigned long long)(__a[1], __b[1]);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_mergel(vector float __a, vector float __b) {
-  return (vector float)(__a[2], __b[2], __a[3], __b[3]);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_mergel(vector double __a, vector double __b) {
-  return (vector double)(__a[1], __b[1]);
-}
-
-/*-- vec_pack ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_pack(vector signed short __a, vector signed short __b) {
-  vector signed char __ac = (vector signed char)__a;
-  vector signed char __bc = (vector signed char)__b;
-  return (vector signed char)(
-    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
-    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_pack(vector bool short __a, vector bool short __b) {
-  vector bool char __ac = (vector bool char)__a;
-  vector bool char __bc = (vector bool char)__b;
-  return (vector bool char)(
-    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
-    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_pack(vector unsigned short __a, vector unsigned short __b) {
-  vector unsigned char __ac = (vector unsigned char)__a;
-  vector unsigned char __bc = (vector unsigned char)__b;
-  return (vector unsigned char)(
-    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
-    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_pack(vector signed int __a, vector signed int __b) {
-  vector signed short __ac = (vector signed short)__a;
-  vector signed short __bc = (vector signed short)__b;
-  return (vector signed short)(
-    __ac[1], __ac[3], __ac[5], __ac[7],
-    __bc[1], __bc[3], __bc[5], __bc[7]);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_pack(vector bool int __a, vector bool int __b) {
-  vector bool short __ac = (vector bool short)__a;
-  vector bool short __bc = (vector bool short)__b;
-  return (vector bool short)(
-    __ac[1], __ac[3], __ac[5], __ac[7],
-    __bc[1], __bc[3], __bc[5], __bc[7]);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_pack(vector unsigned int __a, vector unsigned int __b) {
-  vector unsigned short __ac = (vector unsigned short)__a;
-  vector unsigned short __bc = (vector unsigned short)__b;
-  return (vector unsigned short)(
-    __ac[1], __ac[3], __ac[5], __ac[7],
-    __bc[1], __bc[3], __bc[5], __bc[7]);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_pack(vector signed long long __a, vector signed long long __b) {
-  vector signed int __ac = (vector signed int)__a;
-  vector signed int __bc = (vector signed int)__b;
-  return (vector signed int)(__ac[1], __ac[3], __bc[1], __bc[3]);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_pack(vector bool long long __a, vector bool long long __b) {
-  vector bool int __ac = (vector bool int)__a;
-  vector bool int __bc = (vector bool int)__b;
-  return (vector bool int)(__ac[1], __ac[3], __bc[1], __bc[3]);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_pack(vector unsigned long long __a, vector unsigned long long __b) {
-  vector unsigned int __ac = (vector unsigned int)__a;
-  vector unsigned int __bc = (vector unsigned int)__b;
-  return (vector unsigned int)(__ac[1], __ac[3], __bc[1], __bc[3]);
-}
-
-/*-- vec_packs --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_packs(vector signed short __a, vector signed short __b) {
-  return __builtin_s390_vpksh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_packs(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vpklsh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_packs(vector signed int __a, vector signed int __b) {
-  return __builtin_s390_vpksf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_packs(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vpklsf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_packs(vector signed long long __a, vector signed long long __b) {
-  return __builtin_s390_vpksg(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_packs(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vpklsg(__a, __b);
-}
-
-/*-- vec_packs_cc -----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_packs_cc(vector signed short __a, vector signed short __b, int *__cc) {
-  return __builtin_s390_vpkshs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_packs_cc(vector unsigned short __a, vector unsigned short __b, int *__cc) {
-  return __builtin_s390_vpklshs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_packs_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return __builtin_s390_vpksfs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_packs_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
-  return __builtin_s390_vpklsfs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_packs_cc(vector signed long long __a, vector signed long long __b,
-             int *__cc) {
-  return __builtin_s390_vpksgs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_packs_cc(vector unsigned long long __a, vector unsigned long long __b,
-             int *__cc) {
-  return __builtin_s390_vpklsgs(__a, __b, __cc);
-}
-
-/*-- vec_packsu -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_packsu(vector signed short __a, vector signed short __b) {
-  const vector signed short __zero = (vector signed short)0;
-  return __builtin_s390_vpklsh(
-    (vector unsigned short)(__a >= __zero) & (vector unsigned short)__a,
-    (vector unsigned short)(__b >= __zero) & (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_packsu(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vpklsh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_packsu(vector signed int __a, vector signed int __b) {
-  const vector signed int __zero = (vector signed int)0;
-  return __builtin_s390_vpklsf(
-    (vector unsigned int)(__a >= __zero) & (vector unsigned int)__a,
-    (vector unsigned int)(__b >= __zero) & (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_packsu(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vpklsf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_packsu(vector signed long long __a, vector signed long long __b) {
-  const vector signed long long __zero = (vector signed long long)0;
-  return __builtin_s390_vpklsg(
-    (vector unsigned long long)(__a >= __zero) &
-    (vector unsigned long long)__a,
-    (vector unsigned long long)(__b >= __zero) &
-    (vector unsigned long long)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_packsu(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vpklsg(__a, __b);
-}
-
-/*-- vec_packsu_cc ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_packsu_cc(vector unsigned short __a, vector unsigned short __b, int *__cc) {
-  return __builtin_s390_vpklshs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_packsu_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
-  return __builtin_s390_vpklsfs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_packsu_cc(vector unsigned long long __a, vector unsigned long long __b,
-              int *__cc) {
-  return __builtin_s390_vpklsgs(__a, __b, __cc);
-}
-
-/*-- vec_unpackh ------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_unpackh(vector signed char __a) {
-  return __builtin_s390_vuphb(__a);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_unpackh(vector bool char __a) {
-  return (vector bool short)__builtin_s390_vuphb((vector signed char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_unpackh(vector unsigned char __a) {
-  return __builtin_s390_vuplhb(__a);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_unpackh(vector signed short __a) {
-  return __builtin_s390_vuphh(__a);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_unpackh(vector bool short __a) {
-  return (vector bool int)__builtin_s390_vuphh((vector signed short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_unpackh(vector unsigned short __a) {
-  return __builtin_s390_vuplhh(__a);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_unpackh(vector signed int __a) {
-  return __builtin_s390_vuphf(__a);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_unpackh(vector bool int __a) {
-  return (vector bool long long)__builtin_s390_vuphf((vector signed int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_unpackh(vector unsigned int __a) {
-  return __builtin_s390_vuplhf(__a);
-}
-
-/*-- vec_unpackl ------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_unpackl(vector signed char __a) {
-  return __builtin_s390_vuplb(__a);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_unpackl(vector bool char __a) {
-  return (vector bool short)__builtin_s390_vuplb((vector signed char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_unpackl(vector unsigned char __a) {
-  return __builtin_s390_vupllb(__a);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_unpackl(vector signed short __a) {
-  return __builtin_s390_vuplhw(__a);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_unpackl(vector bool short __a) {
-  return (vector bool int)__builtin_s390_vuplhw((vector signed short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_unpackl(vector unsigned short __a) {
-  return __builtin_s390_vupllh(__a);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_unpackl(vector signed int __a) {
-  return __builtin_s390_vuplf(__a);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_unpackl(vector bool int __a) {
-  return (vector bool long long)__builtin_s390_vuplf((vector signed int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_unpackl(vector unsigned int __a) {
-  return __builtin_s390_vupllf(__a);
-}
-
-/*-- vec_cmpeq --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpeq(vector bool char __a, vector bool char __b) {
-  return (vector bool char)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpeq(vector signed char __a, vector signed char __b) {
-  return (vector bool char)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpeq(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpeq(vector bool short __a, vector bool short __b) {
-  return (vector bool short)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpeq(vector signed short __a, vector signed short __b) {
-  return (vector bool short)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpeq(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpeq(vector bool int __a, vector bool int __b) {
-  return (vector bool int)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpeq(vector signed int __a, vector signed int __b) {
-  return (vector bool int)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpeq(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpeq(vector bool long long __a, vector bool long long __b) {
-  return (vector bool long long)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpeq(vector signed long long __a, vector signed long long __b) {
-  return (vector bool long long)(__a == __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpeq(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector bool long long)(__a == __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool int
-vec_cmpeq(vector float __a, vector float __b) {
-  return (vector bool int)(__a == __b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpeq(vector double __a, vector double __b) {
-  return (vector bool long long)(__a == __b);
-}
-
-/*-- vec_cmpge --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpge(vector signed char __a, vector signed char __b) {
-  return (vector bool char)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpge(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpge(vector signed short __a, vector signed short __b) {
-  return (vector bool short)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpge(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpge(vector signed int __a, vector signed int __b) {
-  return (vector bool int)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpge(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpge(vector signed long long __a, vector signed long long __b) {
-  return (vector bool long long)(__a >= __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpge(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector bool long long)(__a >= __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool int
-vec_cmpge(vector float __a, vector float __b) {
-  return (vector bool int)(__a >= __b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpge(vector double __a, vector double __b) {
-  return (vector bool long long)(__a >= __b);
-}
-
-/*-- vec_cmpgt --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpgt(vector signed char __a, vector signed char __b) {
-  return (vector bool char)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpgt(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpgt(vector signed short __a, vector signed short __b) {
-  return (vector bool short)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpgt(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpgt(vector signed int __a, vector signed int __b) {
-  return (vector bool int)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpgt(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpgt(vector signed long long __a, vector signed long long __b) {
-  return (vector bool long long)(__a > __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpgt(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector bool long long)(__a > __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool int
-vec_cmpgt(vector float __a, vector float __b) {
-  return (vector bool int)(__a > __b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmpgt(vector double __a, vector double __b) {
-  return (vector bool long long)(__a > __b);
-}
-
-/*-- vec_cmple --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmple(vector signed char __a, vector signed char __b) {
-  return (vector bool char)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmple(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmple(vector signed short __a, vector signed short __b) {
-  return (vector bool short)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmple(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmple(vector signed int __a, vector signed int __b) {
-  return (vector bool int)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmple(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmple(vector signed long long __a, vector signed long long __b) {
-  return (vector bool long long)(__a <= __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmple(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector bool long long)(__a <= __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool int
-vec_cmple(vector float __a, vector float __b) {
-  return (vector bool int)(__a <= __b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmple(vector double __a, vector double __b) {
-  return (vector bool long long)(__a <= __b);
-}
-
-/*-- vec_cmplt --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmplt(vector signed char __a, vector signed char __b) {
-  return (vector bool char)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmplt(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmplt(vector signed short __a, vector signed short __b) {
-  return (vector bool short)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmplt(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmplt(vector signed int __a, vector signed int __b) {
-  return (vector bool int)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmplt(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmplt(vector signed long long __a, vector signed long long __b) {
-  return (vector bool long long)(__a < __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmplt(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector bool long long)(__a < __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool int
-vec_cmplt(vector float __a, vector float __b) {
-  return (vector bool int)(__a < __b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector bool long long
-vec_cmplt(vector double __a, vector double __b) {
-  return (vector bool long long)(__a < __b);
-}
-
-/*-- vec_all_eq -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_eq(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfcesbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_eq(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfcedbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_ne -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_ne(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfcesbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_ne(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfcedbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_all_ge -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b,
-                        (vector unsigned char)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b,
-                        (vector unsigned short)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b,
-                        (vector unsigned int)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_ge(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b,
-                        (vector unsigned long long)__a, &__cc);
-  return __cc == 3;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_ge(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_ge(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_gt -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a,
-                        (vector unsigned char)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a,
-                        (vector unsigned short)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a,
-                        (vector unsigned int)__b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_gt(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a,
-                        (vector unsigned long long)__b, &__cc);
-  return __cc == 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_gt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_gt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__a, __b, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_le -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a,
-                        (vector unsigned char)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a,
-                        (vector unsigned short)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a,
-                        (vector unsigned int)__b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
-  return __cc == 3;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_le(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a,
-                        (vector unsigned long long)__b, &__cc);
-  return __cc == 3;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_le(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_le(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_lt -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b,
-                        (vector unsigned char)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b,
-                        (vector unsigned short)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b,
-                        (vector unsigned int)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
-  return __cc == 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
-  return __cc == 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_all_lt(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b,
-                        (vector unsigned long long)__a, &__cc);
-  return __cc == 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_lt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_lt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__b, __a, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_nge ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_nge(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_nge(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_all_ngt ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_ngt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_ngt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__a, __b, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_all_nle ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_nle(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_nle(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_all_nlt ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_nlt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_nlt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__b, __a, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_all_nan ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_nan(vector float __a) {
-  int __cc;
-  __builtin_s390_vftcisb(__a, 15, &__cc);
-  return __cc == 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_nan(vector double __a) {
-  int __cc;
-  __builtin_s390_vftcidb(__a, 15, &__cc);
-  return __cc == 0;
-}
-
-/*-- vec_all_numeric --------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_all_numeric(vector float __a) {
-  int __cc;
-  __builtin_s390_vftcisb(__a, 15, &__cc);
-  return __cc == 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_all_numeric(vector double __a) {
-  int __cc;
-  __builtin_s390_vftcidb(__a, 15, &__cc);
-  return __cc == 3;
-}
-
-/*-- vec_any_eq -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_eq(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfcesbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_eq(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfcedbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-/*-- vec_any_ne -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vceqbs((vector signed char)__a,
-                        (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vceqhs((vector signed short)__a,
-                        (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vceqfs((vector signed int)__a,
-                        (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vceqgs((vector signed long long)__a,
-                        (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_ne(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfcesbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_ne(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfcedbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_any_ge -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b,
-                        (vector unsigned char)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b,
-                        (vector unsigned short)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b,
-                        (vector unsigned int)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_ge(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b,
-                        (vector unsigned long long)__a, &__cc);
-  return __cc != 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_ge(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_ge(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-/*-- vec_any_gt -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a,
-                        (vector unsigned char)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a,
-                        (vector unsigned short)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a,
-                        (vector unsigned int)__b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_gt(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a,
-                        (vector unsigned long long)__b, &__cc);
-  return __cc <= 1;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_gt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_gt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__a, __b, &__cc);
-  return __cc <= 1;
-}
-
-/*-- vec_any_le -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__a,
-                        (vector unsigned char)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__a,
-                        (vector unsigned short)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__a,
-                        (vector unsigned int)__b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
-  return __cc != 0;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_le(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__a,
-                        (vector unsigned long long)__b, &__cc);
-  return __cc != 0;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_le(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_le(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-/*-- vec_any_lt -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool char __a, vector signed char __b) {
-  int __cc;
-  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool char __a, vector unsigned char __b) {
-  int __cc;
-  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool char __a, vector bool char __b) {
-  int __cc;
-  __builtin_s390_vchlbs((vector unsigned char)__b,
-                        (vector unsigned char)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool short __a, vector signed short __b) {
-  int __cc;
-  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool short __a, vector unsigned short __b) {
-  int __cc;
-  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool short __a, vector bool short __b) {
-  int __cc;
-  __builtin_s390_vchlhs((vector unsigned short)__b,
-                        (vector unsigned short)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool int __a, vector signed int __b) {
-  int __cc;
-  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool int __a, vector unsigned int __b) {
-  int __cc;
-  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool int __a, vector bool int __b) {
-  int __cc;
-  __builtin_s390_vchlfs((vector unsigned int)__b,
-                        (vector unsigned int)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector signed long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool long long __a, vector signed long long __b) {
-  int __cc;
-  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
-  return __cc <= 1;
-}
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector unsigned long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool long long __a, vector unsigned long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
-  return __cc <= 1;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai int
-vec_any_lt(vector bool long long __a, vector bool long long __b) {
-  int __cc;
-  __builtin_s390_vchlgs((vector unsigned long long)__b,
-                        (vector unsigned long long)__a, &__cc);
-  return __cc <= 1;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_lt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_lt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__b, __a, &__cc);
-  return __cc <= 1;
-}
-
-/*-- vec_any_nge ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_nge(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_nge(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_any_ngt ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_ngt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_ngt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__a, __b, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_any_nle ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_nle(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchesbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_nle(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchedbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_any_nlt ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_nlt(vector float __a, vector float __b) {
-  int __cc;
-  __builtin_s390_vfchsbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_nlt(vector double __a, vector double __b) {
-  int __cc;
-  __builtin_s390_vfchdbs(__b, __a, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_any_nan ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_nan(vector float __a) {
-  int __cc;
-  __builtin_s390_vftcisb(__a, 15, &__cc);
-  return __cc != 3;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_nan(vector double __a) {
-  int __cc;
-  __builtin_s390_vftcidb(__a, 15, &__cc);
-  return __cc != 3;
-}
-
-/*-- vec_any_numeric --------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_any_numeric(vector float __a) {
-  int __cc;
-  __builtin_s390_vftcisb(__a, 15, &__cc);
-  return __cc != 0;
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_any_numeric(vector double __a) {
-  int __cc;
-  __builtin_s390_vftcidb(__a, 15, &__cc);
-  return __cc != 0;
-}
-
-/*-- vec_andc ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_andc(vector bool char __a, vector bool char __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_andc(vector signed char __a, vector signed char __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_andc(vector bool char __a, vector signed char __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_andc(vector signed char __a, vector bool char __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_andc(vector unsigned char __a, vector unsigned char __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_andc(vector bool char __a, vector unsigned char __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_andc(vector unsigned char __a, vector bool char __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_andc(vector bool short __a, vector bool short __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_andc(vector signed short __a, vector signed short __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_andc(vector bool short __a, vector signed short __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_andc(vector signed short __a, vector bool short __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_andc(vector unsigned short __a, vector unsigned short __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_andc(vector bool short __a, vector unsigned short __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_andc(vector unsigned short __a, vector bool short __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_andc(vector bool int __a, vector bool int __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_andc(vector signed int __a, vector signed int __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_andc(vector bool int __a, vector signed int __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_andc(vector signed int __a, vector bool int __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_andc(vector unsigned int __a, vector unsigned int __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_andc(vector bool int __a, vector unsigned int __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_andc(vector unsigned int __a, vector bool int __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_andc(vector bool long long __a, vector bool long long __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_andc(vector signed long long __a, vector signed long long __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_andc(vector bool long long __a, vector signed long long __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_andc(vector signed long long __a, vector bool long long __b) {
-  return __a & ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_andc(vector unsigned long long __a, vector unsigned long long __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_andc(vector bool long long __a, vector unsigned long long __b) {
-  return __a & ~__b;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_andc(vector unsigned long long __a, vector bool long long __b) {
-  return __a & ~__b;
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_andc(vector float __a, vector float __b) {
-  return (vector float)((vector unsigned int)__a &
-                         ~(vector unsigned int)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_andc(vector double __a, vector double __b) {
-  return (vector double)((vector unsigned long long)__a &
-                         ~(vector unsigned long long)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_andc(vector bool long long __a, vector double __b) {
-  return (vector double)((vector unsigned long long)__a &
-                         ~(vector unsigned long long)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_andc(vector double __a, vector bool long long __b) {
-  return (vector double)((vector unsigned long long)__a &
-                         ~(vector unsigned long long)__b);
-}
-
-/*-- vec_nor ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_nor(vector bool char __a, vector bool char __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_nor(vector signed char __a, vector signed char __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_nor(vector bool char __a, vector signed char __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_nor(vector signed char __a, vector bool char __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_nor(vector unsigned char __a, vector unsigned char __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_nor(vector bool char __a, vector unsigned char __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_nor(vector unsigned char __a, vector bool char __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_nor(vector bool short __a, vector bool short __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_nor(vector signed short __a, vector signed short __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_nor(vector bool short __a, vector signed short __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_nor(vector signed short __a, vector bool short __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_nor(vector unsigned short __a, vector unsigned short __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_nor(vector bool short __a, vector unsigned short __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_nor(vector unsigned short __a, vector bool short __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_nor(vector bool int __a, vector bool int __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_nor(vector signed int __a, vector signed int __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_nor(vector bool int __a, vector signed int __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_nor(vector signed int __a, vector bool int __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_nor(vector unsigned int __a, vector unsigned int __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_nor(vector bool int __a, vector unsigned int __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_nor(vector unsigned int __a, vector bool int __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_nor(vector bool long long __a, vector bool long long __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_nor(vector signed long long __a, vector signed long long __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_nor(vector bool long long __a, vector signed long long __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_nor(vector signed long long __a, vector bool long long __b) {
-  return ~(__a | __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_nor(vector unsigned long long __a, vector unsigned long long __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_nor(vector bool long long __a, vector unsigned long long __b) {
-  return ~(__a | __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_nor(vector unsigned long long __a, vector bool long long __b) {
-  return ~(__a | __b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_nor(vector float __a, vector float __b) {
-  return (vector float)~((vector unsigned int)__a |
-                         (vector unsigned int)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_nor(vector double __a, vector double __b) {
-  return (vector double)~((vector unsigned long long)__a |
-                          (vector unsigned long long)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_nor(vector bool long long __a, vector double __b) {
-  return (vector double)~((vector unsigned long long)__a |
-                          (vector unsigned long long)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_nor(vector double __a, vector bool long long __b) {
-  return (vector double)~((vector unsigned long long)__a |
-                          (vector unsigned long long)__b);
-}
-
-/*-- vec_orc ----------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool char
-vec_orc(vector bool char __a, vector bool char __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_orc(vector signed char __a, vector signed char __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_orc(vector unsigned char __a, vector unsigned char __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_orc(vector bool short __a, vector bool short __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_orc(vector signed short __a, vector signed short __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_orc(vector unsigned short __a, vector unsigned short __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_orc(vector bool int __a, vector bool int __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_orc(vector signed int __a, vector signed int __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_orc(vector unsigned int __a, vector unsigned int __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_orc(vector bool long long __a, vector bool long long __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_orc(vector signed long long __a, vector signed long long __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_orc(vector unsigned long long __a, vector unsigned long long __b) {
-  return __a | ~__b;
-}
-
-static inline __ATTRS_o_ai vector float
-vec_orc(vector float __a, vector float __b) {
-  return (vector float)((vector unsigned int)__a |
-                        ~(vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_orc(vector double __a, vector double __b) {
-  return (vector double)((vector unsigned long long)__a |
-                         ~(vector unsigned long long)__b);
-}
-#endif
-
-/*-- vec_nand ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool char
-vec_nand(vector bool char __a, vector bool char __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_nand(vector signed char __a, vector signed char __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_nand(vector unsigned char __a, vector unsigned char __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_nand(vector bool short __a, vector bool short __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_nand(vector signed short __a, vector signed short __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_nand(vector unsigned short __a, vector unsigned short __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_nand(vector bool int __a, vector bool int __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_nand(vector signed int __a, vector signed int __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_nand(vector unsigned int __a, vector unsigned int __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_nand(vector bool long long __a, vector bool long long __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_nand(vector signed long long __a, vector signed long long __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_nand(vector unsigned long long __a, vector unsigned long long __b) {
-  return ~(__a & __b);
-}
-
-static inline __ATTRS_o_ai vector float
-vec_nand(vector float __a, vector float __b) {
-  return (vector float)~((vector unsigned int)__a &
-                         (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_nand(vector double __a, vector double __b) {
-  return (vector double)~((vector unsigned long long)__a &
-                          (vector unsigned long long)__b);
-}
-#endif
-
-/*-- vec_eqv ----------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector bool char
-vec_eqv(vector bool char __a, vector bool char __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_eqv(vector signed char __a, vector signed char __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_eqv(vector unsigned char __a, vector unsigned char __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_eqv(vector bool short __a, vector bool short __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_eqv(vector signed short __a, vector signed short __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_eqv(vector unsigned short __a, vector unsigned short __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_eqv(vector bool int __a, vector bool int __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_eqv(vector signed int __a, vector signed int __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_eqv(vector unsigned int __a, vector unsigned int __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector bool long long
-vec_eqv(vector bool long long __a, vector bool long long __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_eqv(vector signed long long __a, vector signed long long __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_eqv(vector unsigned long long __a, vector unsigned long long __b) {
-  return ~(__a ^ __b);
-}
-
-static inline __ATTRS_o_ai vector float
-vec_eqv(vector float __a, vector float __b) {
-  return (vector float)~((vector unsigned int)__a ^
-                         (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_eqv(vector double __a, vector double __b) {
-  return (vector double)~((vector unsigned long long)__a ^
-                          (vector unsigned long long)__b);
-}
-#endif
-
-/*-- vec_cntlz --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cntlz(vector signed char __a) {
-  return __builtin_s390_vclzb((vector unsigned char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cntlz(vector unsigned char __a) {
-  return __builtin_s390_vclzb(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cntlz(vector signed short __a) {
-  return __builtin_s390_vclzh((vector unsigned short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cntlz(vector unsigned short __a) {
-  return __builtin_s390_vclzh(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cntlz(vector signed int __a) {
-  return __builtin_s390_vclzf((vector unsigned int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cntlz(vector unsigned int __a) {
-  return __builtin_s390_vclzf(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_cntlz(vector signed long long __a) {
-  return __builtin_s390_vclzg((vector unsigned long long)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_cntlz(vector unsigned long long __a) {
-  return __builtin_s390_vclzg(__a);
-}
-
-/*-- vec_cnttz --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cnttz(vector signed char __a) {
-  return __builtin_s390_vctzb((vector unsigned char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cnttz(vector unsigned char __a) {
-  return __builtin_s390_vctzb(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cnttz(vector signed short __a) {
-  return __builtin_s390_vctzh((vector unsigned short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cnttz(vector unsigned short __a) {
-  return __builtin_s390_vctzh(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cnttz(vector signed int __a) {
-  return __builtin_s390_vctzf((vector unsigned int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cnttz(vector unsigned int __a) {
-  return __builtin_s390_vctzf(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_cnttz(vector signed long long __a) {
-  return __builtin_s390_vctzg((vector unsigned long long)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_cnttz(vector unsigned long long __a) {
-  return __builtin_s390_vctzg(__a);
-}
-
-/*-- vec_popcnt -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_popcnt(vector signed char __a) {
-  return __builtin_s390_vpopctb((vector unsigned char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_popcnt(vector unsigned char __a) {
-  return __builtin_s390_vpopctb(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_popcnt(vector signed short __a) {
-  return __builtin_s390_vpopcth((vector unsigned short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_popcnt(vector unsigned short __a) {
-  return __builtin_s390_vpopcth(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_popcnt(vector signed int __a) {
-  return __builtin_s390_vpopctf((vector unsigned int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_popcnt(vector unsigned int __a) {
-  return __builtin_s390_vpopctf(__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_popcnt(vector signed long long __a) {
-  return __builtin_s390_vpopctg((vector unsigned long long)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_popcnt(vector unsigned long long __a) {
-  return __builtin_s390_vpopctg(__a);
-}
-
-/*-- vec_rl -----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_rl(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_verllvb(
-    (vector unsigned char)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_rl(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_verllvb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_rl(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_verllvh(
-    (vector unsigned short)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_rl(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_verllvh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_rl(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_verllvf(
-    (vector unsigned int)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_rl(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_verllvf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_rl(vector signed long long __a, vector unsigned long long __b) {
-  return (vector signed long long)__builtin_s390_verllvg(
-    (vector unsigned long long)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_rl(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_verllvg(__a, __b);
-}
-
-/*-- vec_rli ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_rli(vector signed char __a, unsigned long __b) {
-  return (vector signed char)__builtin_s390_verllb(
-    (vector unsigned char)__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_rli(vector unsigned char __a, unsigned long __b) {
-  return __builtin_s390_verllb(__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_rli(vector signed short __a, unsigned long __b) {
-  return (vector signed short)__builtin_s390_verllh(
-    (vector unsigned short)__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_rli(vector unsigned short __a, unsigned long __b) {
-  return __builtin_s390_verllh(__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_rli(vector signed int __a, unsigned long __b) {
-  return (vector signed int)__builtin_s390_verllf(
-    (vector unsigned int)__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_rli(vector unsigned int __a, unsigned long __b) {
-  return __builtin_s390_verllf(__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_rli(vector signed long long __a, unsigned long __b) {
-  return (vector signed long long)__builtin_s390_verllg(
-    (vector unsigned long long)__a, (int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_rli(vector unsigned long long __a, unsigned long __b) {
-  return __builtin_s390_verllg(__a, (int)__b);
-}
-
-/*-- vec_rl_mask ------------------------------------------------------------*/
-
-extern __ATTRS_o vector signed char
-vec_rl_mask(vector signed char __a, vector unsigned char __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector unsigned char
-vec_rl_mask(vector unsigned char __a, vector unsigned char __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector signed short
-vec_rl_mask(vector signed short __a, vector unsigned short __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector unsigned short
-vec_rl_mask(vector unsigned short __a, vector unsigned short __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector signed int
-vec_rl_mask(vector signed int __a, vector unsigned int __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector unsigned int
-vec_rl_mask(vector unsigned int __a, vector unsigned int __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector signed long long
-vec_rl_mask(vector signed long long __a, vector unsigned long long __b,
-            unsigned char __c) __constant(__c);
-
-extern __ATTRS_o vector unsigned long long
-vec_rl_mask(vector unsigned long long __a, vector unsigned long long __b,
-            unsigned char __c) __constant(__c);
-
-#define vec_rl_mask(X, Y, Z) ((__typeof__((vec_rl_mask)((X), (Y), (Z)))) \
-  __extension__ ({ \
-    vector unsigned char __res; \
-    vector unsigned char __x = (vector unsigned char)(X); \
-    vector unsigned char __y = (vector unsigned char)(Y); \
-    switch (sizeof ((X)[0])) { \
-    case 1: __res = (vector unsigned char) __builtin_s390_verimb( \
-             (vector unsigned char)__x, (vector unsigned char)__x, \
-             (vector unsigned char)__y, (Z)); break; \
-    case 2: __res = (vector unsigned char) __builtin_s390_verimh( \
-             (vector unsigned short)__x, (vector unsigned short)__x, \
-             (vector unsigned short)__y, (Z)); break; \
-    case 4: __res = (vector unsigned char) __builtin_s390_verimf( \
-             (vector unsigned int)__x, (vector unsigned int)__x, \
-             (vector unsigned int)__y, (Z)); break; \
-    default: __res = (vector unsigned char) __builtin_s390_verimg( \
-             (vector unsigned long long)__x, (vector unsigned long long)__x, \
-             (vector unsigned long long)__y, (Z)); break; \
-    } __res; }))
-
-/*-- vec_sll ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_sll(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_sll(vector signed char __a, vector unsigned short __b) {
-  return (vector signed char)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_sll(vector signed char __a, vector unsigned int __b) {
-  return (vector signed char)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sll(vector bool char __a, vector unsigned char __b) {
-  return (vector bool char)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sll(vector bool char __a, vector unsigned short __b) {
-  return (vector bool char)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sll(vector bool char __a, vector unsigned int __b) {
-  return (vector bool char)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sll(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsl(__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_sll(vector unsigned char __a, vector unsigned short __b) {
-  return __builtin_s390_vsl(__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_sll(vector unsigned char __a, vector unsigned int __b) {
-  return __builtin_s390_vsl(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_sll(vector signed short __a, vector unsigned char __b) {
-  return (vector signed short)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_sll(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_sll(vector signed short __a, vector unsigned int __b) {
-  return (vector signed short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sll(vector bool short __a, vector unsigned char __b) {
-  return (vector bool short)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sll(vector bool short __a, vector unsigned short __b) {
-  return (vector bool short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sll(vector bool short __a, vector unsigned int __b) {
-  return (vector bool short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_sll(vector unsigned short __a, vector unsigned char __b) {
-  return (vector unsigned short)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_sll(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_sll(vector unsigned short __a, vector unsigned int __b) {
-  return (vector unsigned short)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_sll(vector signed int __a, vector unsigned char __b) {
-  return (vector signed int)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_sll(vector signed int __a, vector unsigned short __b) {
-  return (vector signed int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_sll(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sll(vector bool int __a, vector unsigned char __b) {
-  return (vector bool int)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sll(vector bool int __a, vector unsigned short __b) {
-  return (vector bool int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sll(vector bool int __a, vector unsigned int __b) {
-  return (vector bool int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sll(vector unsigned int __a, vector unsigned char __b) {
-  return (vector unsigned int)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_sll(vector unsigned int __a, vector unsigned short __b) {
-  return (vector unsigned int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_sll(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_sll(vector signed long long __a, vector unsigned char __b) {
-  return (vector signed long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_sll(vector signed long long __a, vector unsigned short __b) {
-  return (vector signed long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_sll(vector signed long long __a, vector unsigned int __b) {
-  return (vector signed long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sll(vector bool long long __a, vector unsigned char __b) {
-  return (vector bool long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sll(vector bool long long __a, vector unsigned short __b) {
-  return (vector bool long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sll(vector bool long long __a, vector unsigned int __b) {
-  return (vector bool long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sll(vector unsigned long long __a, vector unsigned char __b) {
-  return (vector unsigned long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sll(vector unsigned long long __a, vector unsigned short __b) {
-  return (vector unsigned long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sll(vector unsigned long long __a, vector unsigned int __b) {
-  return (vector unsigned long long)__builtin_s390_vsl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_slb ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_slb(vector signed char __a, vector signed char __b) {
-  return (vector signed char)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_slb(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vslb(
-    (vector unsigned char)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_slb(vector unsigned char __a, vector signed char __b) {
-  return __builtin_s390_vslb(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_slb(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vslb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_slb(vector signed short __a, vector signed short __b) {
-  return (vector signed short)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_slb(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_slb(vector unsigned short __a, vector signed short __b) {
-  return (vector unsigned short)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_slb(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_slb(vector signed int __a, vector signed int __b) {
-  return (vector signed int)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_slb(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_slb(vector unsigned int __a, vector signed int __b) {
-  return (vector unsigned int)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_slb(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_slb(vector signed long long __a, vector signed long long __b) {
-  return (vector signed long long)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_slb(vector signed long long __a, vector unsigned long long __b) {
-  return (vector signed long long)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_slb(vector unsigned long long __a, vector signed long long __b) {
-  return (vector unsigned long long)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_slb(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector unsigned long long)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_slb(vector float __a, vector signed int __b) {
-  return (vector float)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector float
-vec_slb(vector float __a, vector unsigned int __b) {
-  return (vector float)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_slb(vector double __a, vector signed long long __b) {
-  return (vector double)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_slb(vector double __a, vector unsigned long long __b) {
-  return (vector double)__builtin_s390_vslb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_sld ----------------------------------------------------------------*/
-
-extern __ATTRS_o vector signed char
-vec_sld(vector signed char __a, vector signed char __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector bool char
-vec_sld(vector bool char __a, vector bool char __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector unsigned char
-vec_sld(vector unsigned char __a, vector unsigned char __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector signed short
-vec_sld(vector signed short __a, vector signed short __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector bool short
-vec_sld(vector bool short __a, vector bool short __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector unsigned short
-vec_sld(vector unsigned short __a, vector unsigned short __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector signed int
-vec_sld(vector signed int __a, vector signed int __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector bool int
-vec_sld(vector bool int __a, vector bool int __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector unsigned int
-vec_sld(vector unsigned int __a, vector unsigned int __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector signed long long
-vec_sld(vector signed long long __a, vector signed long long __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector bool long long
-vec_sld(vector bool long long __a, vector bool long long __b, int __c)
-  __constant_range(__c, 0, 15);
-
-extern __ATTRS_o vector unsigned long long
-vec_sld(vector unsigned long long __a, vector unsigned long long __b, int __c)
-  __constant_range(__c, 0, 15);
-
-#if __ARCH__ >= 12
-extern __ATTRS_o vector float
-vec_sld(vector float __a, vector float __b, int __c)
-  __constant_range(__c, 0, 15);
-#endif
-
-extern __ATTRS_o vector double
-vec_sld(vector double __a, vector double __b, int __c)
-  __constant_range(__c, 0, 15);
-
-#define vec_sld(X, Y, Z) ((__typeof__((vec_sld)((X), (Y), (Z)))) \
-  __builtin_s390_vsldb((vector unsigned char)(X), \
-                       (vector unsigned char)(Y), (Z)))
-
-/*-- vec_sldw ---------------------------------------------------------------*/
-
-extern __ATTRS_o vector signed char
-vec_sldw(vector signed char __a, vector signed char __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector unsigned char
-vec_sldw(vector unsigned char __a, vector unsigned char __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector signed short
-vec_sldw(vector signed short __a, vector signed short __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector unsigned short
-vec_sldw(vector unsigned short __a, vector unsigned short __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector signed int
-vec_sldw(vector signed int __a, vector signed int __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector unsigned int
-vec_sldw(vector unsigned int __a, vector unsigned int __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector signed long long
-vec_sldw(vector signed long long __a, vector signed long long __b, int __c)
-  __constant_range(__c, 0, 3);
-
-extern __ATTRS_o vector unsigned long long
-vec_sldw(vector unsigned long long __a, vector unsigned long long __b, int __c)
-  __constant_range(__c, 0, 3);
-
-// This prototype is deprecated.
-extern __ATTRS_o vector double
-vec_sldw(vector double __a, vector double __b, int __c)
-  __constant_range(__c, 0, 3);
-
-#define vec_sldw(X, Y, Z) ((__typeof__((vec_sldw)((X), (Y), (Z)))) \
-  __builtin_s390_vsldb((vector unsigned char)(X), \
-                       (vector unsigned char)(Y), (Z) * 4))
-
-/*-- vec_sral ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_sral(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_sral(vector signed char __a, vector unsigned short __b) {
-  return (vector signed char)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_sral(vector signed char __a, vector unsigned int __b) {
-  return (vector signed char)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sral(vector bool char __a, vector unsigned char __b) {
-  return (vector bool char)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sral(vector bool char __a, vector unsigned short __b) {
-  return (vector bool char)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_sral(vector bool char __a, vector unsigned int __b) {
-  return (vector bool char)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sral(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsra(__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_sral(vector unsigned char __a, vector unsigned short __b) {
-  return __builtin_s390_vsra(__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_sral(vector unsigned char __a, vector unsigned int __b) {
-  return __builtin_s390_vsra(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_sral(vector signed short __a, vector unsigned char __b) {
-  return (vector signed short)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_sral(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_sral(vector signed short __a, vector unsigned int __b) {
-  return (vector signed short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sral(vector bool short __a, vector unsigned char __b) {
-  return (vector bool short)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sral(vector bool short __a, vector unsigned short __b) {
-  return (vector bool short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_sral(vector bool short __a, vector unsigned int __b) {
-  return (vector bool short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_sral(vector unsigned short __a, vector unsigned char __b) {
-  return (vector unsigned short)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_sral(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_sral(vector unsigned short __a, vector unsigned int __b) {
-  return (vector unsigned short)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_sral(vector signed int __a, vector unsigned char __b) {
-  return (vector signed int)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_sral(vector signed int __a, vector unsigned short __b) {
-  return (vector signed int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_sral(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sral(vector bool int __a, vector unsigned char __b) {
-  return (vector bool int)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sral(vector bool int __a, vector unsigned short __b) {
-  return (vector bool int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_sral(vector bool int __a, vector unsigned int __b) {
-  return (vector bool int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sral(vector unsigned int __a, vector unsigned char __b) {
-  return (vector unsigned int)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_sral(vector unsigned int __a, vector unsigned short __b) {
-  return (vector unsigned int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_sral(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_sral(vector signed long long __a, vector unsigned char __b) {
-  return (vector signed long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_sral(vector signed long long __a, vector unsigned short __b) {
-  return (vector signed long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_sral(vector signed long long __a, vector unsigned int __b) {
-  return (vector signed long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sral(vector bool long long __a, vector unsigned char __b) {
-  return (vector bool long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sral(vector bool long long __a, vector unsigned short __b) {
-  return (vector bool long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_sral(vector bool long long __a, vector unsigned int __b) {
-  return (vector bool long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sral(vector unsigned long long __a, vector unsigned char __b) {
-  return (vector unsigned long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sral(vector unsigned long long __a, vector unsigned short __b) {
-  return (vector unsigned long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sral(vector unsigned long long __a, vector unsigned int __b) {
-  return (vector unsigned long long)__builtin_s390_vsra(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_srab ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_srab(vector signed char __a, vector signed char __b) {
-  return (vector signed char)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_srab(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vsrab(
-    (vector unsigned char)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_srab(vector unsigned char __a, vector signed char __b) {
-  return __builtin_s390_vsrab(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_srab(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsrab(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_srab(vector signed short __a, vector signed short __b) {
-  return (vector signed short)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_srab(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_srab(vector unsigned short __a, vector signed short __b) {
-  return (vector unsigned short)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_srab(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_srab(vector signed int __a, vector signed int __b) {
-  return (vector signed int)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_srab(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_srab(vector unsigned int __a, vector signed int __b) {
-  return (vector unsigned int)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_srab(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_srab(vector signed long long __a, vector signed long long __b) {
-  return (vector signed long long)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_srab(vector signed long long __a, vector unsigned long long __b) {
-  return (vector signed long long)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srab(vector unsigned long long __a, vector signed long long __b) {
-  return (vector unsigned long long)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srab(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector unsigned long long)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_srab(vector float __a, vector signed int __b) {
-  return (vector float)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector float
-vec_srab(vector float __a, vector unsigned int __b) {
-  return (vector float)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_srab(vector double __a, vector signed long long __b) {
-  return (vector double)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_srab(vector double __a, vector unsigned long long __b) {
-  return (vector double)__builtin_s390_vsrab(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_srl ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_srl(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_srl(vector signed char __a, vector unsigned short __b) {
-  return (vector signed char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_srl(vector signed char __a, vector unsigned int __b) {
-  return (vector signed char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_srl(vector bool char __a, vector unsigned char __b) {
-  return (vector bool char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_srl(vector bool char __a, vector unsigned short __b) {
-  return (vector bool char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool char
-vec_srl(vector bool char __a, vector unsigned int __b) {
-  return (vector bool char)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_srl(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsrl(__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_srl(vector unsigned char __a, vector unsigned short __b) {
-  return __builtin_s390_vsrl(__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_srl(vector unsigned char __a, vector unsigned int __b) {
-  return __builtin_s390_vsrl(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_srl(vector signed short __a, vector unsigned char __b) {
-  return (vector signed short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_srl(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_srl(vector signed short __a, vector unsigned int __b) {
-  return (vector signed short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_srl(vector bool short __a, vector unsigned char __b) {
-  return (vector bool short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_srl(vector bool short __a, vector unsigned short __b) {
-  return (vector bool short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool short
-vec_srl(vector bool short __a, vector unsigned int __b) {
-  return (vector bool short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_srl(vector unsigned short __a, vector unsigned char __b) {
-  return (vector unsigned short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_srl(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_srl(vector unsigned short __a, vector unsigned int __b) {
-  return (vector unsigned short)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_srl(vector signed int __a, vector unsigned char __b) {
-  return (vector signed int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_srl(vector signed int __a, vector unsigned short __b) {
-  return (vector signed int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_srl(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_srl(vector bool int __a, vector unsigned char __b) {
-  return (vector bool int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_srl(vector bool int __a, vector unsigned short __b) {
-  return (vector bool int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool int
-vec_srl(vector bool int __a, vector unsigned int __b) {
-  return (vector bool int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_srl(vector unsigned int __a, vector unsigned char __b) {
-  return (vector unsigned int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_srl(vector unsigned int __a, vector unsigned short __b) {
-  return (vector unsigned int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_srl(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_srl(vector signed long long __a, vector unsigned char __b) {
-  return (vector signed long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_srl(vector signed long long __a, vector unsigned short __b) {
-  return (vector signed long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_srl(vector signed long long __a, vector unsigned int __b) {
-  return (vector signed long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_srl(vector bool long long __a, vector unsigned char __b) {
-  return (vector bool long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_srl(vector bool long long __a, vector unsigned short __b) {
-  return (vector bool long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector bool long long
-vec_srl(vector bool long long __a, vector unsigned int __b) {
-  return (vector bool long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srl(vector unsigned long long __a, vector unsigned char __b) {
-  return (vector unsigned long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, __b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srl(vector unsigned long long __a, vector unsigned short __b) {
-  return (vector unsigned long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srl(vector unsigned long long __a, vector unsigned int __b) {
-  return (vector unsigned long long)__builtin_s390_vsrl(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_srb ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_srb(vector signed char __a, vector signed char __b) {
-  return (vector signed char)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_srb(vector signed char __a, vector unsigned char __b) {
-  return (vector signed char)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_srb(vector unsigned char __a, vector signed char __b) {
-  return __builtin_s390_vsrlb(__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_srb(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsrlb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_srb(vector signed short __a, vector signed short __b) {
-  return (vector signed short)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_srb(vector signed short __a, vector unsigned short __b) {
-  return (vector signed short)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_srb(vector unsigned short __a, vector signed short __b) {
-  return (vector unsigned short)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_srb(vector unsigned short __a, vector unsigned short __b) {
-  return (vector unsigned short)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_srb(vector signed int __a, vector signed int __b) {
-  return (vector signed int)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_srb(vector signed int __a, vector unsigned int __b) {
-  return (vector signed int)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_srb(vector unsigned int __a, vector signed int __b) {
-  return (vector unsigned int)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_srb(vector unsigned int __a, vector unsigned int __b) {
-  return (vector unsigned int)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_srb(vector signed long long __a, vector signed long long __b) {
-  return (vector signed long long)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_srb(vector signed long long __a, vector unsigned long long __b) {
-  return (vector signed long long)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srb(vector unsigned long long __a, vector signed long long __b) {
-  return (vector unsigned long long)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_srb(vector unsigned long long __a, vector unsigned long long __b) {
-  return (vector unsigned long long)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_srb(vector float __a, vector signed int __b) {
-  return (vector float)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector float
-vec_srb(vector float __a, vector unsigned int __b) {
-  return (vector float)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_srb(vector double __a, vector signed long long __b) {
-  return (vector double)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_srb(vector double __a, vector unsigned long long __b) {
-  return (vector double)__builtin_s390_vsrlb(
-    (vector unsigned char)__a, (vector unsigned char)__b);
-}
-
-/*-- vec_abs ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_abs(vector signed char __a) {
-  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed char)0));
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_abs(vector signed short __a) {
-  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed short)0));
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_abs(vector signed int __a) {
-  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed int)0));
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_abs(vector signed long long __a) {
-  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed long long)0));
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_abs(vector float __a) {
-  return __builtin_s390_vflpsb(__a);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_abs(vector double __a) {
-  return __builtin_s390_vflpdb(__a);
-}
-
-/*-- vec_nabs ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_nabs(vector float __a) {
-  return __builtin_s390_vflnsb(__a);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_nabs(vector double __a) {
-  return __builtin_s390_vflndb(__a);
-}
-
-/*-- vec_max ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_max(vector signed char __a, vector signed char __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_max(vector signed char __a, vector bool char __b) {
-  vector signed char __bc = (vector signed char)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_max(vector bool char __a, vector signed char __b) {
-  vector signed char __ac = (vector signed char)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_max(vector unsigned char __a, vector unsigned char __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_max(vector unsigned char __a, vector bool char __b) {
-  vector unsigned char __bc = (vector unsigned char)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_max(vector bool char __a, vector unsigned char __b) {
-  vector unsigned char __ac = (vector unsigned char)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_max(vector signed short __a, vector signed short __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_max(vector signed short __a, vector bool short __b) {
-  vector signed short __bc = (vector signed short)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_max(vector bool short __a, vector signed short __b) {
-  vector signed short __ac = (vector signed short)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_max(vector unsigned short __a, vector unsigned short __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_max(vector unsigned short __a, vector bool short __b) {
-  vector unsigned short __bc = (vector unsigned short)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_max(vector bool short __a, vector unsigned short __b) {
-  vector unsigned short __ac = (vector unsigned short)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_max(vector signed int __a, vector signed int __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_max(vector signed int __a, vector bool int __b) {
-  vector signed int __bc = (vector signed int)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_max(vector bool int __a, vector signed int __b) {
-  vector signed int __ac = (vector signed int)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_max(vector unsigned int __a, vector unsigned int __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_max(vector unsigned int __a, vector bool int __b) {
-  vector unsigned int __bc = (vector unsigned int)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_max(vector bool int __a, vector unsigned int __b) {
-  vector unsigned int __ac = (vector unsigned int)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_max(vector signed long long __a, vector signed long long __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_max(vector signed long long __a, vector bool long long __b) {
-  vector signed long long __bc = (vector signed long long)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_max(vector bool long long __a, vector signed long long __b) {
-  vector signed long long __ac = (vector signed long long)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_max(vector unsigned long long __a, vector unsigned long long __b) {
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_max(vector unsigned long long __a, vector bool long long __b) {
-  vector unsigned long long __bc = (vector unsigned long long)__b;
-  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_max(vector bool long long __a, vector unsigned long long __b) {
-  vector unsigned long long __ac = (vector unsigned long long)__a;
-  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_max(vector float __a, vector float __b) {
-  return __builtin_s390_vfmaxsb(__a, __b, 0);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_max(vector double __a, vector double __b) {
-#if __ARCH__ >= 12
-  return __builtin_s390_vfmaxdb(__a, __b, 0);
-#else
-  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
-#endif
-}
-
-/*-- vec_min ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_min(vector signed char __a, vector signed char __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_min(vector signed char __a, vector bool char __b) {
-  vector signed char __bc = (vector signed char)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed char
-vec_min(vector bool char __a, vector signed char __b) {
-  vector signed char __ac = (vector signed char)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_min(vector unsigned char __a, vector unsigned char __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_min(vector unsigned char __a, vector bool char __b) {
-  vector unsigned char __bc = (vector unsigned char)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned char
-vec_min(vector bool char __a, vector unsigned char __b) {
-  vector unsigned char __ac = (vector unsigned char)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_min(vector signed short __a, vector signed short __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_min(vector signed short __a, vector bool short __b) {
-  vector signed short __bc = (vector signed short)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed short
-vec_min(vector bool short __a, vector signed short __b) {
-  vector signed short __ac = (vector signed short)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_min(vector unsigned short __a, vector unsigned short __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_min(vector unsigned short __a, vector bool short __b) {
-  vector unsigned short __bc = (vector unsigned short)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned short
-vec_min(vector bool short __a, vector unsigned short __b) {
-  vector unsigned short __ac = (vector unsigned short)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_min(vector signed int __a, vector signed int __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_min(vector signed int __a, vector bool int __b) {
-  vector signed int __bc = (vector signed int)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed int
-vec_min(vector bool int __a, vector signed int __b) {
-  vector signed int __ac = (vector signed int)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_min(vector unsigned int __a, vector unsigned int __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_min(vector unsigned int __a, vector bool int __b) {
-  vector unsigned int __bc = (vector unsigned int)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned int
-vec_min(vector bool int __a, vector unsigned int __b) {
-  vector unsigned int __ac = (vector unsigned int)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_min(vector signed long long __a, vector signed long long __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_min(vector signed long long __a, vector bool long long __b) {
-  vector signed long long __bc = (vector signed long long)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_min(vector bool long long __a, vector signed long long __b) {
-  vector signed long long __ac = (vector signed long long)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_min(vector unsigned long long __a, vector unsigned long long __b) {
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_min(vector unsigned long long __a, vector bool long long __b) {
-  vector unsigned long long __bc = (vector unsigned long long)__b;
-  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_min(vector bool long long __a, vector unsigned long long __b) {
-  vector unsigned long long __ac = (vector unsigned long long)__a;
-  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_min(vector float __a, vector float __b) {
-  return __builtin_s390_vfminsb(__a, __b, 0);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_min(vector double __a, vector double __b) {
-#if __ARCH__ >= 12
-  return __builtin_s390_vfmindb(__a, __b, 0);
-#else
-  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
-#endif
-}
-
-/*-- vec_add_u128 -----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_add_u128(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vaq(__a, __b);
-}
-
-/*-- vec_addc ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_addc(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vaccb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_addc(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vacch(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_addc(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vaccf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_addc(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vaccg(__a, __b);
-}
-
-/*-- vec_addc_u128 ----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_addc_u128(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vaccq(__a, __b);
-}
-
-/*-- vec_adde_u128 ----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_adde_u128(vector unsigned char __a, vector unsigned char __b,
-              vector unsigned char __c) {
-  return __builtin_s390_vacq(__a, __b, __c);
-}
-
-/*-- vec_addec_u128 ---------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_addec_u128(vector unsigned char __a, vector unsigned char __b,
-               vector unsigned char __c) {
-  return __builtin_s390_vacccq(__a, __b, __c);
-}
-
-/*-- vec_avg ----------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_avg(vector signed char __a, vector signed char __b) {
-  return __builtin_s390_vavgb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_avg(vector signed short __a, vector signed short __b) {
-  return __builtin_s390_vavgh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_avg(vector signed int __a, vector signed int __b) {
-  return __builtin_s390_vavgf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_avg(vector signed long long __a, vector signed long long __b) {
-  return __builtin_s390_vavgg(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_avg(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vavglb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_avg(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vavglh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_avg(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vavglf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_avg(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vavglg(__a, __b);
-}
-
-/*-- vec_checksum -----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned int
-vec_checksum(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vcksm(__a, __b);
-}
-
-/*-- vec_gfmsum -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_gfmsum(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vgfmb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_gfmsum(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vgfmh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_gfmsum(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vgfmf(__a, __b);
-}
-
-/*-- vec_gfmsum_128 ---------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_gfmsum_128(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vgfmg(__a, __b);
-}
-
-/*-- vec_gfmsum_accum -------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_gfmsum_accum(vector unsigned char __a, vector unsigned char __b,
-                 vector unsigned short __c) {
-  return __builtin_s390_vgfmab(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_gfmsum_accum(vector unsigned short __a, vector unsigned short __b,
-                 vector unsigned int __c) {
-  return __builtin_s390_vgfmah(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_gfmsum_accum(vector unsigned int __a, vector unsigned int __b,
-                 vector unsigned long long __c) {
-  return __builtin_s390_vgfmaf(__a, __b, __c);
-}
-
-/*-- vec_gfmsum_accum_128 ---------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_gfmsum_accum_128(vector unsigned long long __a,
-                     vector unsigned long long __b,
-                     vector unsigned char __c) {
-  return __builtin_s390_vgfmag(__a, __b, __c);
-}
-
-/*-- vec_mladd --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_mladd(vector signed char __a, vector signed char __b,
-          vector signed char __c) {
-  return __a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_mladd(vector unsigned char __a, vector signed char __b,
-          vector signed char __c) {
-  return (vector signed char)__a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed char
-vec_mladd(vector signed char __a, vector unsigned char __b,
-          vector unsigned char __c) {
-  return __a * (vector signed char)__b + (vector signed char)__c;
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_mladd(vector unsigned char __a, vector unsigned char __b,
-          vector unsigned char __c) {
-  return __a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mladd(vector signed short __a, vector signed short __b,
-          vector signed short __c) {
-  return __a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mladd(vector unsigned short __a, vector signed short __b,
-          vector signed short __c) {
-  return (vector signed short)__a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mladd(vector signed short __a, vector unsigned short __b,
-          vector unsigned short __c) {
-  return __a * (vector signed short)__b + (vector signed short)__c;
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mladd(vector unsigned short __a, vector unsigned short __b,
-          vector unsigned short __c) {
-  return __a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mladd(vector signed int __a, vector signed int __b,
-          vector signed int __c) {
-  return __a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mladd(vector unsigned int __a, vector signed int __b,
-          vector signed int __c) {
-  return (vector signed int)__a * __b + __c;
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mladd(vector signed int __a, vector unsigned int __b,
-          vector unsigned int __c) {
-  return __a * (vector signed int)__b + (vector signed int)__c;
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mladd(vector unsigned int __a, vector unsigned int __b,
-          vector unsigned int __c) {
-  return __a * __b + __c;
-}
-
-/*-- vec_mhadd --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_mhadd(vector signed char __a, vector signed char __b,
-          vector signed char __c) {
-  return __builtin_s390_vmahb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_mhadd(vector unsigned char __a, vector unsigned char __b,
-          vector unsigned char __c) {
-  return __builtin_s390_vmalhb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mhadd(vector signed short __a, vector signed short __b,
-          vector signed short __c) {
-  return __builtin_s390_vmahh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mhadd(vector unsigned short __a, vector unsigned short __b,
-          vector unsigned short __c) {
-  return __builtin_s390_vmalhh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mhadd(vector signed int __a, vector signed int __b,
-          vector signed int __c) {
-  return __builtin_s390_vmahf(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mhadd(vector unsigned int __a, vector unsigned int __b,
-          vector unsigned int __c) {
-  return __builtin_s390_vmalhf(__a, __b, __c);
-}
-
-/*-- vec_meadd --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_meadd(vector signed char __a, vector signed char __b,
-          vector signed short __c) {
-  return __builtin_s390_vmaeb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_meadd(vector unsigned char __a, vector unsigned char __b,
-          vector unsigned short __c) {
-  return __builtin_s390_vmaleb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_meadd(vector signed short __a, vector signed short __b,
-          vector signed int __c) {
-  return __builtin_s390_vmaeh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_meadd(vector unsigned short __a, vector unsigned short __b,
-          vector unsigned int __c) {
-  return __builtin_s390_vmaleh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_meadd(vector signed int __a, vector signed int __b,
-          vector signed long long __c) {
-  return __builtin_s390_vmaef(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_meadd(vector unsigned int __a, vector unsigned int __b,
-          vector unsigned long long __c) {
-  return __builtin_s390_vmalef(__a, __b, __c);
-}
-
-/*-- vec_moadd --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_moadd(vector signed char __a, vector signed char __b,
-          vector signed short __c) {
-  return __builtin_s390_vmaob(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_moadd(vector unsigned char __a, vector unsigned char __b,
-          vector unsigned short __c) {
-  return __builtin_s390_vmalob(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_moadd(vector signed short __a, vector signed short __b,
-          vector signed int __c) {
-  return __builtin_s390_vmaoh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_moadd(vector unsigned short __a, vector unsigned short __b,
-          vector unsigned int __c) {
-  return __builtin_s390_vmaloh(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_moadd(vector signed int __a, vector signed int __b,
-          vector signed long long __c) {
-  return __builtin_s390_vmaof(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_moadd(vector unsigned int __a, vector unsigned int __b,
-          vector unsigned long long __c) {
-  return __builtin_s390_vmalof(__a, __b, __c);
-}
-
-/*-- vec_mulh ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_mulh(vector signed char __a, vector signed char __b) {
-  return __builtin_s390_vmhb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_mulh(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vmlhb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_mulh(vector signed short __a, vector signed short __b) {
-  return __builtin_s390_vmhh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mulh(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vmlhh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mulh(vector signed int __a, vector signed int __b) {
-  return __builtin_s390_vmhf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mulh(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vmlhf(__a, __b);
-}
-
-/*-- vec_mule ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_mule(vector signed char __a, vector signed char __b) {
-  return __builtin_s390_vmeb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mule(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vmleb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mule(vector signed short __a, vector signed short __b) {
-  return __builtin_s390_vmeh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mule(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vmleh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_mule(vector signed int __a, vector signed int __b) {
-  return __builtin_s390_vmef(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_mule(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vmlef(__a, __b);
-}
-
-/*-- vec_mulo ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed short
-vec_mulo(vector signed char __a, vector signed char __b) {
-  return __builtin_s390_vmob(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_mulo(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vmlob(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_mulo(vector signed short __a, vector signed short __b) {
-  return __builtin_s390_vmoh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_mulo(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vmloh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed long long
-vec_mulo(vector signed int __a, vector signed int __b) {
-  return __builtin_s390_vmof(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_mulo(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vmlof(__a, __b);
-}
-
-/*-- vec_msum_u128 ----------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-#define vec_msum_u128(X, Y, Z, W) \
-  ((vector unsigned char)__builtin_s390_vmslg((X), (Y), (Z), (W)));
-#endif
-
-/*-- vec_sub_u128 -----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_sub_u128(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsq(__a, __b);
-}
-
-/*-- vec_subc ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_subc(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vscbib(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_subc(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vscbih(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_subc(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vscbif(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_subc(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vscbig(__a, __b);
-}
-
-/*-- vec_subc_u128 ----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_subc_u128(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vscbiq(__a, __b);
-}
-
-/*-- vec_sube_u128 ----------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_sube_u128(vector unsigned char __a, vector unsigned char __b,
-              vector unsigned char __c) {
-  return __builtin_s390_vsbiq(__a, __b, __c);
-}
-
-/*-- vec_subec_u128 ---------------------------------------------------------*/
-
-static inline __ATTRS_ai vector unsigned char
-vec_subec_u128(vector unsigned char __a, vector unsigned char __b,
-               vector unsigned char __c) {
-  return __builtin_s390_vsbcbiq(__a, __b, __c);
-}
-
-/*-- vec_sum2 ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sum2(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vsumgh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_sum2(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vsumgf(__a, __b);
-}
-
-/*-- vec_sum_u128 -----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sum_u128(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vsumqf(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_sum_u128(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vsumqg(__a, __b);
-}
-
-/*-- vec_sum4 ---------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sum4(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vsumb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_sum4(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vsumh(__a, __b);
-}
-
-/*-- vec_test_mask ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector signed char __a, vector unsigned char __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vtm(__a, __b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector signed short __a, vector unsigned short __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector signed int __a, vector unsigned int __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector signed long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector unsigned long long __a, vector unsigned long long __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai int
-vec_test_mask(vector float __a, vector unsigned int __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-#endif
-
-static inline __ATTRS_o_ai int
-vec_test_mask(vector double __a, vector unsigned long long __b) {
-  return __builtin_s390_vtm((vector unsigned char)__a,
-                            (vector unsigned char)__b);
-}
-
-/*-- vec_madd ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_madd(vector float __a, vector float __b, vector float __c) {
-  return __builtin_s390_vfmasb(__a, __b, __c);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_madd(vector double __a, vector double __b, vector double __c) {
-  return __builtin_s390_vfmadb(__a, __b, __c);
-}
-
-/*-- vec_msub ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_msub(vector float __a, vector float __b, vector float __c) {
-  return __builtin_s390_vfmssb(__a, __b, __c);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_msub(vector double __a, vector double __b, vector double __c) {
-  return __builtin_s390_vfmsdb(__a, __b, __c);
-}
-
-/*-- vec_nmadd ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_nmadd(vector float __a, vector float __b, vector float __c) {
-  return __builtin_s390_vfnmasb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_nmadd(vector double __a, vector double __b, vector double __c) {
-  return __builtin_s390_vfnmadb(__a, __b, __c);
-}
-#endif
-
-/*-- vec_nmsub ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_nmsub(vector float __a, vector float __b, vector float __c) {
-  return __builtin_s390_vfnmssb(__a, __b, __c);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_nmsub(vector double __a, vector double __b, vector double __c) {
-  return __builtin_s390_vfnmsdb(__a, __b, __c);
-}
-#endif
-
-/*-- vec_sqrt ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_sqrt(vector float __a) {
-  return __builtin_s390_vfsqsb(__a);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_sqrt(vector double __a) {
-  return __builtin_s390_vfsqdb(__a);
-}
-
-/*-- vec_ld2f ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_ai vector double
-vec_ld2f(const float *__ptr) {
-  typedef float __v2f32 __attribute__((__vector_size__(8)));
-  return __builtin_convertvector(*(const __v2f32 *)__ptr, vector double);
-}
-
-/*-- vec_st2f ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_ai void
-vec_st2f(vector double __a, float *__ptr) {
-  typedef float __v2f32 __attribute__((__vector_size__(8)));
-  *(__v2f32 *)__ptr = __builtin_convertvector(__a, __v2f32);
-}
-
-/*-- vec_ctd ----------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_ctd(vector signed long long __a, int __b)
-  __constant_range(__b, 0, 31) {
-  vector double __conv = __builtin_convertvector(__a, vector double);
-  __conv *= (vector double)(vector unsigned long long)((0x3ffULL - __b) << 52);
-  return __conv;
-}
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector double
-vec_ctd(vector unsigned long long __a, int __b)
-  __constant_range(__b, 0, 31) {
-  vector double __conv = __builtin_convertvector(__a, vector double);
-  __conv *= (vector double)(vector unsigned long long)((0x3ffULL - __b) << 52);
-  return __conv;
-}
-
-/*-- vec_ctsl ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector signed long long
-vec_ctsl(vector double __a, int __b)
-  __constant_range(__b, 0, 31) {
-  __a *= (vector double)(vector unsigned long long)((0x3ffULL + __b) << 52);
-  return __builtin_convertvector(__a, vector signed long long);
-}
-
-/*-- vec_ctul ---------------------------------------------------------------*/
-
-// This prototype is deprecated.
-static inline __ATTRS_o_ai vector unsigned long long
-vec_ctul(vector double __a, int __b)
-  __constant_range(__b, 0, 31) {
-  __a *= (vector double)(vector unsigned long long)((0x3ffULL + __b) << 52);
-  return __builtin_convertvector(__a, vector unsigned long long);
-}
-
-/*-- vec_doublee ------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_ai vector double
-vec_doublee(vector float __a) {
-  typedef float __v2f32 __attribute__((__vector_size__(8)));
-  __v2f32 __pack = __builtin_shufflevector(__a, __a, 0, 2);
-  return __builtin_convertvector(__pack, vector double);
-}
-#endif
-
-/*-- vec_floate -------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_ai vector float
-vec_floate(vector double __a) {
-  typedef float __v2f32 __attribute__((__vector_size__(8)));
-  __v2f32 __pack = __builtin_convertvector(__a, __v2f32);
-  return __builtin_shufflevector(__pack, __pack, 0, -1, 1, -1);
-}
-#endif
-
-/*-- vec_double -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector double
-vec_double(vector signed long long __a) {
-  return __builtin_convertvector(__a, vector double);
-}
-
-static inline __ATTRS_o_ai vector double
-vec_double(vector unsigned long long __a) {
-  return __builtin_convertvector(__a, vector double);
-}
-
-/*-- vec_signed -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed long long
-vec_signed(vector double __a) {
-  return __builtin_convertvector(__a, vector signed long long);
-}
-
-/*-- vec_unsigned -----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned long long
-vec_unsigned(vector double __a) {
-  return __builtin_convertvector(__a, vector unsigned long long);
-}
-
-/*-- vec_roundp -------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_roundp(vector float __a) {
-  return __builtin_s390_vfisb(__a, 4, 6);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_roundp(vector double __a) {
-  return __builtin_s390_vfidb(__a, 4, 6);
-}
-
-/*-- vec_ceil ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_ceil(vector float __a) {
-  // On this platform, vec_ceil never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfisb(__a, 4, 6);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_ceil(vector double __a) {
-  // On this platform, vec_ceil never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfidb(__a, 4, 6);
-}
-
-/*-- vec_roundm -------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_roundm(vector float __a) {
-  return __builtin_s390_vfisb(__a, 4, 7);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_roundm(vector double __a) {
-  return __builtin_s390_vfidb(__a, 4, 7);
-}
-
-/*-- vec_floor --------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_floor(vector float __a) {
-  // On this platform, vec_floor never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfisb(__a, 4, 7);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_floor(vector double __a) {
-  // On this platform, vec_floor never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfidb(__a, 4, 7);
-}
-
-/*-- vec_roundz -------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_roundz(vector float __a) {
-  return __builtin_s390_vfisb(__a, 4, 5);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_roundz(vector double __a) {
-  return __builtin_s390_vfidb(__a, 4, 5);
-}
-
-/*-- vec_trunc --------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_trunc(vector float __a) {
-  // On this platform, vec_trunc never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfisb(__a, 4, 5);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_trunc(vector double __a) {
-  // On this platform, vec_trunc never triggers the IEEE-inexact exception.
-  return __builtin_s390_vfidb(__a, 4, 5);
-}
-
-/*-- vec_roundc -------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_roundc(vector float __a) {
-  return __builtin_s390_vfisb(__a, 4, 0);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_roundc(vector double __a) {
-  return __builtin_s390_vfidb(__a, 4, 0);
-}
-
-/*-- vec_rint ---------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_rint(vector float __a) {
-  // vec_rint may trigger the IEEE-inexact exception.
-  return __builtin_s390_vfisb(__a, 0, 0);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_rint(vector double __a) {
-  // vec_rint may trigger the IEEE-inexact exception.
-  return __builtin_s390_vfidb(__a, 0, 0);
-}
-
-/*-- vec_round --------------------------------------------------------------*/
-
-#if __ARCH__ >= 12
-static inline __ATTRS_o_ai vector float
-vec_round(vector float __a) {
-  return __builtin_s390_vfisb(__a, 4, 4);
-}
-#endif
-
-static inline __ATTRS_o_ai vector double
-vec_round(vector double __a) {
-  return __builtin_s390_vfidb(__a, 4, 4);
-}
-
-/*-- vec_fp_test_data_class -------------------------------------------------*/
-
-#if __ARCH__ >= 12
-extern __ATTRS_o vector bool int
-vec_fp_test_data_class(vector float __a, int __b, int *__c)
-  __constant_range(__b, 0, 4095);
-
-extern __ATTRS_o vector bool long long
-vec_fp_test_data_class(vector double __a, int __b, int *__c)
-  __constant_range(__b, 0, 4095);
-
-#define vec_fp_test_data_class(X, Y, Z) \
-  ((__typeof__((vec_fp_test_data_class)((X), (Y), (Z)))) \
-   __extension__ ({ \
-     vector unsigned char __res; \
-     vector unsigned char __x = (vector unsigned char)(X); \
-     int *__z = (Z); \
-     switch (sizeof ((X)[0])) { \
-     case 4:  __res = (vector unsigned char) \
-                      __builtin_s390_vftcisb((vector float)__x, (Y), __z); \
-              break; \
-     default: __res = (vector unsigned char) \
-                      __builtin_s390_vftcidb((vector double)__x, (Y), __z); \
-              break; \
-     } __res; }))
-#else
-#define vec_fp_test_data_class(X, Y, Z) \
-  ((vector bool long long)__builtin_s390_vftcidb((X), (Y), (Z)))
-#endif
-
-#define __VEC_CLASS_FP_ZERO_P (1 << 11)
-#define __VEC_CLASS_FP_ZERO_N (1 << 10)
-#define __VEC_CLASS_FP_ZERO (__VEC_CLASS_FP_ZERO_P | __VEC_CLASS_FP_ZERO_N)
-#define __VEC_CLASS_FP_NORMAL_P (1 << 9)
-#define __VEC_CLASS_FP_NORMAL_N (1 << 8)
-#define __VEC_CLASS_FP_NORMAL (__VEC_CLASS_FP_NORMAL_P | \
-                               __VEC_CLASS_FP_NORMAL_N)
-#define __VEC_CLASS_FP_SUBNORMAL_P (1 << 7)
-#define __VEC_CLASS_FP_SUBNORMAL_N (1 << 6)
-#define __VEC_CLASS_FP_SUBNORMAL (__VEC_CLASS_FP_SUBNORMAL_P | \
-                                  __VEC_CLASS_FP_SUBNORMAL_N)
-#define __VEC_CLASS_FP_INFINITY_P (1 << 5)
-#define __VEC_CLASS_FP_INFINITY_N (1 << 4)
-#define __VEC_CLASS_FP_INFINITY (__VEC_CLASS_FP_INFINITY_P | \
-                                 __VEC_CLASS_FP_INFINITY_N)
-#define __VEC_CLASS_FP_QNAN_P (1 << 3)
-#define __VEC_CLASS_FP_QNAN_N (1 << 2)
-#define __VEC_CLASS_FP_QNAN (__VEC_CLASS_FP_QNAN_P | __VEC_CLASS_FP_QNAN_N)
-#define __VEC_CLASS_FP_SNAN_P (1 << 1)
-#define __VEC_CLASS_FP_SNAN_N (1 << 0)
-#define __VEC_CLASS_FP_SNAN (__VEC_CLASS_FP_SNAN_P | __VEC_CLASS_FP_SNAN_N)
-#define __VEC_CLASS_FP_NAN (__VEC_CLASS_FP_QNAN | __VEC_CLASS_FP_SNAN)
-#define __VEC_CLASS_FP_NOT_NORMAL (__VEC_CLASS_FP_NAN | \
-                                   __VEC_CLASS_FP_SUBNORMAL | \
-                                   __VEC_CLASS_FP_ZERO | \
-                                   __VEC_CLASS_FP_INFINITY)
-
-/*-- vec_cp_until_zero ------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cp_until_zero(vector signed char __a) {
-  return (vector signed char)__builtin_s390_vistrb((vector unsigned char)__a);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cp_until_zero(vector bool char __a) {
-  return (vector bool char)__builtin_s390_vistrb((vector unsigned char)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cp_until_zero(vector unsigned char __a) {
-  return __builtin_s390_vistrb(__a);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cp_until_zero(vector signed short __a) {
-  return (vector signed short)__builtin_s390_vistrh((vector unsigned short)__a);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cp_until_zero(vector bool short __a) {
-  return (vector bool short)__builtin_s390_vistrh((vector unsigned short)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cp_until_zero(vector unsigned short __a) {
-  return __builtin_s390_vistrh(__a);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cp_until_zero(vector signed int __a) {
-  return (vector signed int)__builtin_s390_vistrf((vector unsigned int)__a);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cp_until_zero(vector bool int __a) {
-  return (vector bool int)__builtin_s390_vistrf((vector unsigned int)__a);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cp_until_zero(vector unsigned int __a) {
-  return __builtin_s390_vistrf(__a);
-}
-
-/*-- vec_cp_until_zero_cc ---------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cp_until_zero_cc(vector signed char __a, int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vistrbs((vector unsigned char)__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_cp_until_zero_cc(vector bool char __a, int *__cc) {
-  return (vector bool char)
-    __builtin_s390_vistrbs((vector unsigned char)__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cp_until_zero_cc(vector unsigned char __a, int *__cc) {
-  return __builtin_s390_vistrbs(__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cp_until_zero_cc(vector signed short __a, int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vistrhs((vector unsigned short)__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cp_until_zero_cc(vector bool short __a, int *__cc) {
-  return (vector bool short)
-    __builtin_s390_vistrhs((vector unsigned short)__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cp_until_zero_cc(vector unsigned short __a, int *__cc) {
-  return __builtin_s390_vistrhs(__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cp_until_zero_cc(vector signed int __a, int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vistrfs((vector unsigned int)__a, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cp_until_zero_cc(vector bool int __a, int *__cc) {
-  return (vector bool int)__builtin_s390_vistrfs((vector unsigned int)__a,
-                                                 __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cp_until_zero_cc(vector unsigned int __a, int *__cc) {
-  return __builtin_s390_vistrfs(__a, __cc);
-}
-
-/*-- vec_cmpeq_idx ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpeq_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfeeb((vector unsigned char)__a,
-                         (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfeeb((vector unsigned char)__a,
-                              (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfeeb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpeq_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfeeh((vector unsigned short)__a,
-                         (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfeeh((vector unsigned short)__a,
-                              (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfeeh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpeq_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfeef((vector unsigned int)__a,
-                         (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfeef((vector unsigned int)__a,
-                              (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfeef(__a, __b);
-}
-
-/*-- vec_cmpeq_idx_cc -------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpeq_idx_cc(vector signed char __a, vector signed char __b, int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfeebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfeebs((vector unsigned char)__a,
-                               (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                 int *__cc) {
-  return __builtin_s390_vfeebs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpeq_idx_cc(vector signed short __a, vector signed short __b, int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfeehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return __builtin_s390_vfeehs((vector unsigned short)__a,
-                               (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                 int *__cc) {
-  return __builtin_s390_vfeehs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpeq_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfeefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfeefs((vector unsigned int)__a,
-                               (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_idx_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
-  return __builtin_s390_vfeefs(__a, __b, __cc);
-}
-
-/*-- vec_cmpeq_or_0_idx -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpeq_or_0_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfeezb((vector unsigned char)__a,
-                          (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_or_0_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfeezb((vector unsigned char)__a,
-                               (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfeezb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpeq_or_0_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfeezh((vector unsigned short)__a,
-                          (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_or_0_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfeezh((vector unsigned short)__a,
-                               (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfeezh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpeq_or_0_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfeezf((vector unsigned int)__a,
-                          (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_or_0_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfeezf((vector unsigned int)__a,
-                               (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfeezf(__a, __b);
-}
-
-/*-- vec_cmpeq_or_0_idx_cc --------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpeq_or_0_idx_cc(vector signed char __a, vector signed char __b,
-                      int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfeezbs((vector unsigned char)__a,
-                           (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_or_0_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfeezbs((vector unsigned char)__a,
-                                (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpeq_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                      int *__cc) {
-  return __builtin_s390_vfeezbs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpeq_or_0_idx_cc(vector signed short __a, vector signed short __b,
-                      int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfeezhs((vector unsigned short)__a,
-                           (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_or_0_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return __builtin_s390_vfeezhs((vector unsigned short)__a,
-                                (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpeq_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                      int *__cc) {
-  return __builtin_s390_vfeezhs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpeq_or_0_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfeezfs((vector unsigned int)__a,
-                           (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_or_0_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfeezfs((vector unsigned int)__a,
-                                (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpeq_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                      int *__cc) {
-  return __builtin_s390_vfeezfs(__a, __b, __cc);
-}
-
-/*-- vec_cmpne_idx ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpne_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfeneb((vector unsigned char)__a,
-                          (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfeneb((vector unsigned char)__a,
-                               (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfeneb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpne_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfeneh((vector unsigned short)__a,
-                          (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfeneh((vector unsigned short)__a,
-                               (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfeneh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpne_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfenef((vector unsigned int)__a,
-                          (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfenef((vector unsigned int)__a,
-                               (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfenef(__a, __b);
-}
-
-/*-- vec_cmpne_idx_cc -------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpne_idx_cc(vector signed char __a, vector signed char __b, int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfenebs((vector unsigned char)__a,
-                           (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfenebs((vector unsigned char)__a,
-                                (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                 int *__cc) {
-  return __builtin_s390_vfenebs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpne_idx_cc(vector signed short __a, vector signed short __b, int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfenehs((vector unsigned short)__a,
-                           (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return __builtin_s390_vfenehs((vector unsigned short)__a,
-                                (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                 int *__cc) {
-  return __builtin_s390_vfenehs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpne_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfenefs((vector unsigned int)__a,
-                           (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfenefs((vector unsigned int)__a,
-                                (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_idx_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
-  return __builtin_s390_vfenefs(__a, __b, __cc);
-}
-
-/*-- vec_cmpne_or_0_idx -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpne_or_0_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfenezb((vector unsigned char)__a,
-                           (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_or_0_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfenezb((vector unsigned char)__a,
-                                (vector unsigned char)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfenezb(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpne_or_0_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfenezh((vector unsigned short)__a,
-                           (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_or_0_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfenezh((vector unsigned short)__a,
-                                (vector unsigned short)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfenezh(__a, __b);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpne_or_0_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfenezf((vector unsigned int)__a,
-                           (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_or_0_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfenezf((vector unsigned int)__a,
-                                (vector unsigned int)__b);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfenezf(__a, __b);
-}
-
-/*-- vec_cmpne_or_0_idx_cc --------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_cmpne_or_0_idx_cc(vector signed char __a, vector signed char __b,
-                      int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfenezbs((vector unsigned char)__a,
-                            (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_or_0_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfenezbs((vector unsigned char)__a,
-                                 (vector unsigned char)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpne_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                      int *__cc) {
-  return __builtin_s390_vfenezbs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_cmpne_or_0_idx_cc(vector signed short __a, vector signed short __b,
-                      int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfenezhs((vector unsigned short)__a,
-                            (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_or_0_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return __builtin_s390_vfenezhs((vector unsigned short)__a,
-                                 (vector unsigned short)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpne_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                      int *__cc) {
-  return __builtin_s390_vfenezhs(__a, __b, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_cmpne_or_0_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfenezfs((vector unsigned int)__a,
-                            (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_or_0_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfenezfs((vector unsigned int)__a,
-                                 (vector unsigned int)__b, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpne_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                      int *__cc) {
-  return __builtin_s390_vfenezfs(__a, __b, __cc);
-}
-
-/*-- vec_cmprg --------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmprg(vector unsigned char __a, vector unsigned char __b,
-          vector unsigned char __c) {
-  return (vector bool char)__builtin_s390_vstrcb(__a, __b, __c, 4);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmprg(vector unsigned short __a, vector unsigned short __b,
-          vector unsigned short __c) {
-  return (vector bool short)__builtin_s390_vstrch(__a, __b, __c, 4);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmprg(vector unsigned int __a, vector unsigned int __b,
-          vector unsigned int __c) {
-  return (vector bool int)__builtin_s390_vstrcf(__a, __b, __c, 4);
-}
-
-/*-- vec_cmprg_cc -----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmprg_cc(vector unsigned char __a, vector unsigned char __b,
-             vector unsigned char __c, int *__cc) {
-  return (vector bool char)__builtin_s390_vstrcbs(__a, __b, __c, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmprg_cc(vector unsigned short __a, vector unsigned short __b,
-             vector unsigned short __c, int *__cc) {
-  return (vector bool short)__builtin_s390_vstrchs(__a, __b, __c, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmprg_cc(vector unsigned int __a, vector unsigned int __b,
-             vector unsigned int __c, int *__cc) {
-  return (vector bool int)__builtin_s390_vstrcfs(__a, __b, __c, 4, __cc);
-}
-
-/*-- vec_cmprg_idx ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmprg_idx(vector unsigned char __a, vector unsigned char __b,
-              vector unsigned char __c) {
-  return __builtin_s390_vstrcb(__a, __b, __c, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmprg_idx(vector unsigned short __a, vector unsigned short __b,
-              vector unsigned short __c) {
-  return __builtin_s390_vstrch(__a, __b, __c, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmprg_idx(vector unsigned int __a, vector unsigned int __b,
-              vector unsigned int __c) {
-  return __builtin_s390_vstrcf(__a, __b, __c, 0);
-}
-
-/*-- vec_cmprg_idx_cc -------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmprg_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                 vector unsigned char __c, int *__cc) {
-  return __builtin_s390_vstrcbs(__a, __b, __c, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmprg_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                 vector unsigned short __c, int *__cc) {
-  return __builtin_s390_vstrchs(__a, __b, __c, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmprg_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                 vector unsigned int __c, int *__cc) {
-  return __builtin_s390_vstrcfs(__a, __b, __c, 0, __cc);
-}
-
-/*-- vec_cmprg_or_0_idx -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmprg_or_0_idx(vector unsigned char __a, vector unsigned char __b,
-                   vector unsigned char __c) {
-  return __builtin_s390_vstrczb(__a, __b, __c, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmprg_or_0_idx(vector unsigned short __a, vector unsigned short __b,
-                   vector unsigned short __c) {
-  return __builtin_s390_vstrczh(__a, __b, __c, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmprg_or_0_idx(vector unsigned int __a, vector unsigned int __b,
-                   vector unsigned int __c) {
-  return __builtin_s390_vstrczf(__a, __b, __c, 0);
-}
-
-/*-- vec_cmprg_or_0_idx_cc --------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmprg_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                      vector unsigned char __c, int *__cc) {
-  return __builtin_s390_vstrczbs(__a, __b, __c, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmprg_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                      vector unsigned short __c, int *__cc) {
-  return __builtin_s390_vstrczhs(__a, __b, __c, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmprg_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                      vector unsigned int __c, int *__cc) {
-  return __builtin_s390_vstrczfs(__a, __b, __c, 0, __cc);
-}
-
-/*-- vec_cmpnrg -------------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpnrg(vector unsigned char __a, vector unsigned char __b,
-           vector unsigned char __c) {
-  return (vector bool char)__builtin_s390_vstrcb(__a, __b, __c, 12);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpnrg(vector unsigned short __a, vector unsigned short __b,
-           vector unsigned short __c) {
-  return (vector bool short)__builtin_s390_vstrch(__a, __b, __c, 12);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpnrg(vector unsigned int __a, vector unsigned int __b,
-           vector unsigned int __c) {
-  return (vector bool int)__builtin_s390_vstrcf(__a, __b, __c, 12);
-}
-
-/*-- vec_cmpnrg_cc ----------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_cmpnrg_cc(vector unsigned char __a, vector unsigned char __b,
-              vector unsigned char __c, int *__cc) {
-  return (vector bool char)__builtin_s390_vstrcbs(__a, __b, __c, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_cmpnrg_cc(vector unsigned short __a, vector unsigned short __b,
-              vector unsigned short __c, int *__cc) {
-  return (vector bool short)__builtin_s390_vstrchs(__a, __b, __c, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_cmpnrg_cc(vector unsigned int __a, vector unsigned int __b,
-              vector unsigned int __c, int *__cc) {
-  return (vector bool int)__builtin_s390_vstrcfs(__a, __b, __c, 12, __cc);
-}
-
-/*-- vec_cmpnrg_idx ---------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpnrg_idx(vector unsigned char __a, vector unsigned char __b,
-               vector unsigned char __c) {
-  return __builtin_s390_vstrcb(__a, __b, __c, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpnrg_idx(vector unsigned short __a, vector unsigned short __b,
-               vector unsigned short __c) {
-  return __builtin_s390_vstrch(__a, __b, __c, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpnrg_idx(vector unsigned int __a, vector unsigned int __b,
-               vector unsigned int __c) {
-  return __builtin_s390_vstrcf(__a, __b, __c, 8);
-}
-
-/*-- vec_cmpnrg_idx_cc ------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpnrg_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                  vector unsigned char __c, int *__cc) {
-  return __builtin_s390_vstrcbs(__a, __b, __c, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpnrg_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                  vector unsigned short __c, int *__cc) {
-  return __builtin_s390_vstrchs(__a, __b, __c, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpnrg_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                  vector unsigned int __c, int *__cc) {
-  return __builtin_s390_vstrcfs(__a, __b, __c, 8, __cc);
-}
-
-/*-- vec_cmpnrg_or_0_idx ----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpnrg_or_0_idx(vector unsigned char __a, vector unsigned char __b,
-                    vector unsigned char __c) {
-  return __builtin_s390_vstrczb(__a, __b, __c, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpnrg_or_0_idx(vector unsigned short __a, vector unsigned short __b,
-                    vector unsigned short __c) {
-  return __builtin_s390_vstrczh(__a, __b, __c, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpnrg_or_0_idx(vector unsigned int __a, vector unsigned int __b,
-                    vector unsigned int __c) {
-  return __builtin_s390_vstrczf(__a, __b, __c, 8);
-}
-
-/*-- vec_cmpnrg_or_0_idx_cc -------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_cmpnrg_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                       vector unsigned char __c, int *__cc) {
-  return __builtin_s390_vstrczbs(__a, __b, __c, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_cmpnrg_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                       vector unsigned short __c, int *__cc) {
-  return __builtin_s390_vstrczhs(__a, __b, __c, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_cmpnrg_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                       vector unsigned int __c, int *__cc) {
-  return __builtin_s390_vstrczfs(__a, __b, __c, 8, __cc);
-}
-
-/*-- vec_find_any_eq --------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq(vector signed char __a, vector signed char __b) {
-  return (vector bool char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq(vector bool char __a, vector bool char __b) {
-  return (vector bool char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)__builtin_s390_vfaeb(__a, __b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq(vector signed short __a, vector signed short __b) {
-  return (vector bool short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq(vector bool short __a, vector bool short __b) {
-  return (vector bool short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)__builtin_s390_vfaeh(__a, __b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq(vector signed int __a, vector signed int __b) {
-  return (vector bool int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq(vector bool int __a, vector bool int __b) {
-  return (vector bool int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 4);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)__builtin_s390_vfaef(__a, __b, 4);
-}
-
-/*-- vec_find_any_eq_cc -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq_cc(vector signed char __a, vector signed char __b, int *__cc) {
-  return (vector bool char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return (vector bool char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_eq_cc(vector unsigned char __a, vector unsigned char __b,
-                   int *__cc) {
-  return (vector bool char)__builtin_s390_vfaebs(__a, __b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq_cc(vector signed short __a, vector signed short __b,
-                   int *__cc) {
-  return (vector bool short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return (vector bool short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_eq_cc(vector unsigned short __a, vector unsigned short __b,
-                   int *__cc) {
-  return (vector bool short)__builtin_s390_vfaehs(__a, __b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector bool int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return (vector bool int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 4, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_eq_cc(vector unsigned int __a, vector unsigned int __b,
-                   int *__cc) {
-  return (vector bool int)__builtin_s390_vfaefs(__a, __b, 4, __cc);
-}
-
-/*-- vec_find_any_eq_idx ----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_eq_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfaeb((vector unsigned char)__a,
-                              (vector unsigned char)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfaeb(__a, __b, 0);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_eq_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfaeh((vector unsigned short)__a,
-                              (vector unsigned short)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfaeh(__a, __b, 0);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_eq_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfaef((vector unsigned int)__a,
-                              (vector unsigned int)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfaef(__a, __b, 0);
-}
-
-/*-- vec_find_any_eq_idx_cc -------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_eq_idx_cc(vector signed char __a, vector signed char __b,
-                       int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfaebs((vector unsigned char)__a,
-                               (vector unsigned char)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                       int *__cc) {
-  return __builtin_s390_vfaebs(__a, __b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_eq_idx_cc(vector signed short __a, vector signed short __b,
-                       int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_idx_cc(vector bool short __a, vector bool short __b,
-                       int *__cc) {
-  return __builtin_s390_vfaehs((vector unsigned short)__a,
-                               (vector unsigned short)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                       int *__cc) {
-  return __builtin_s390_vfaehs(__a, __b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_eq_idx_cc(vector signed int __a, vector signed int __b,
-                       int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfaefs((vector unsigned int)__a,
-                               (vector unsigned int)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                       int *__cc) {
-  return __builtin_s390_vfaefs(__a, __b, 0, __cc);
-}
-
-/*-- vec_find_any_eq_or_0_idx -----------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_eq_or_0_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfaezb((vector unsigned char)__a,
-                          (vector unsigned char)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_or_0_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfaezb((vector unsigned char)__a,
-                               (vector unsigned char)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfaezb(__a, __b, 0);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_eq_or_0_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfaezh((vector unsigned short)__a,
-                          (vector unsigned short)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_or_0_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfaezh((vector unsigned short)__a,
-                               (vector unsigned short)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfaezh(__a, __b, 0);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_eq_or_0_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfaezf((vector unsigned int)__a,
-                          (vector unsigned int)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_or_0_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfaezf((vector unsigned int)__a,
-                               (vector unsigned int)__b, 0);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfaezf(__a, __b, 0);
-}
-
-/*-- vec_find_any_eq_or_0_idx_cc --------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_eq_or_0_idx_cc(vector signed char __a, vector signed char __b,
-                            int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfaezbs((vector unsigned char)__a,
-                           (vector unsigned char)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_or_0_idx_cc(vector bool char __a, vector bool char __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezbs((vector unsigned char)__a,
-                                (vector unsigned char)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_eq_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezbs(__a, __b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_eq_or_0_idx_cc(vector signed short __a, vector signed short __b,
-                            int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfaezhs((vector unsigned short)__a,
-                           (vector unsigned short)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_or_0_idx_cc(vector bool short __a, vector bool short __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezhs((vector unsigned short)__a,
-                                (vector unsigned short)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_eq_or_0_idx_cc(vector unsigned short __a,
-                            vector unsigned short __b, int *__cc) {
-  return __builtin_s390_vfaezhs(__a, __b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_eq_or_0_idx_cc(vector signed int __a, vector signed int __b,
-                            int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfaezfs((vector unsigned int)__a,
-                           (vector unsigned int)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_or_0_idx_cc(vector bool int __a, vector bool int __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezfs((vector unsigned int)__a,
-                                (vector unsigned int)__b, 0, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_eq_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezfs(__a, __b, 0, __cc);
-}
-
-/*-- vec_find_any_ne --------------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne(vector signed char __a, vector signed char __b) {
-  return (vector bool char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne(vector bool char __a, vector bool char __b) {
-  return (vector bool char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne(vector unsigned char __a, vector unsigned char __b) {
-  return (vector bool char)__builtin_s390_vfaeb(__a, __b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne(vector signed short __a, vector signed short __b) {
-  return (vector bool short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne(vector bool short __a, vector bool short __b) {
-  return (vector bool short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne(vector unsigned short __a, vector unsigned short __b) {
-  return (vector bool short)__builtin_s390_vfaeh(__a, __b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne(vector signed int __a, vector signed int __b) {
-  return (vector bool int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne(vector bool int __a, vector bool int __b) {
-  return (vector bool int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 12);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne(vector unsigned int __a, vector unsigned int __b) {
-  return (vector bool int)__builtin_s390_vfaef(__a, __b, 12);
-}
-
-/*-- vec_find_any_ne_cc -----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne_cc(vector signed char __a, vector signed char __b, int *__cc) {
-  return (vector bool char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return (vector bool char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool char
-vec_find_any_ne_cc(vector unsigned char __a, vector unsigned char __b,
-                   int *__cc) {
-  return (vector bool char)__builtin_s390_vfaebs(__a, __b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne_cc(vector signed short __a, vector signed short __b,
-                   int *__cc) {
-  return (vector bool short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne_cc(vector bool short __a, vector bool short __b, int *__cc) {
-  return (vector bool short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool short
-vec_find_any_ne_cc(vector unsigned short __a, vector unsigned short __b,
-                   int *__cc) {
-  return (vector bool short)__builtin_s390_vfaehs(__a, __b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne_cc(vector signed int __a, vector signed int __b, int *__cc) {
-  return (vector bool int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return (vector bool int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 12, __cc);
-}
-
-static inline __ATTRS_o_ai vector bool int
-vec_find_any_ne_cc(vector unsigned int __a, vector unsigned int __b,
-                   int *__cc) {
-  return (vector bool int)__builtin_s390_vfaefs(__a, __b, 12, __cc);
-}
-
-/*-- vec_find_any_ne_idx ----------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_ne_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfaeb((vector unsigned char)__a,
-                         (vector unsigned char)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfaeb((vector unsigned char)__a,
-                              (vector unsigned char)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfaeb(__a, __b, 8);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_ne_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfaeh((vector unsigned short)__a,
-                         (vector unsigned short)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfaeh((vector unsigned short)__a,
-                              (vector unsigned short)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfaeh(__a, __b, 8);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_ne_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfaef((vector unsigned int)__a,
-                         (vector unsigned int)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfaef((vector unsigned int)__a,
-                              (vector unsigned int)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfaef(__a, __b, 8);
-}
-
-/*-- vec_find_any_ne_idx_cc -------------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_ne_idx_cc(vector signed char __a, vector signed char __b,
-                       int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfaebs((vector unsigned char)__a,
-                          (vector unsigned char)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
-  return __builtin_s390_vfaebs((vector unsigned char)__a,
-                               (vector unsigned char)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                       int *__cc) {
-  return __builtin_s390_vfaebs(__a, __b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_ne_idx_cc(vector signed short __a, vector signed short __b,
-                       int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfaehs((vector unsigned short)__a,
-                          (vector unsigned short)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_idx_cc(vector bool short __a, vector bool short __b,
-                       int *__cc) {
-  return __builtin_s390_vfaehs((vector unsigned short)__a,
-                               (vector unsigned short)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_idx_cc(vector unsigned short __a, vector unsigned short __b,
-                       int *__cc) {
-  return __builtin_s390_vfaehs(__a, __b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_ne_idx_cc(vector signed int __a, vector signed int __b,
-                       int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfaefs((vector unsigned int)__a,
-                          (vector unsigned int)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
-  return __builtin_s390_vfaefs((vector unsigned int)__a,
-                               (vector unsigned int)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                       int *__cc) {
-  return __builtin_s390_vfaefs(__a, __b, 8, __cc);
-}
-
-/*-- vec_find_any_ne_or_0_idx -----------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_ne_or_0_idx(vector signed char __a, vector signed char __b) {
-  return (vector signed char)
-    __builtin_s390_vfaezb((vector unsigned char)__a,
-                          (vector unsigned char)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_or_0_idx(vector bool char __a, vector bool char __b) {
-  return __builtin_s390_vfaezb((vector unsigned char)__a,
-                               (vector unsigned char)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
-  return __builtin_s390_vfaezb(__a, __b, 8);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_ne_or_0_idx(vector signed short __a, vector signed short __b) {
-  return (vector signed short)
-    __builtin_s390_vfaezh((vector unsigned short)__a,
-                          (vector unsigned short)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_or_0_idx(vector bool short __a, vector bool short __b) {
-  return __builtin_s390_vfaezh((vector unsigned short)__a,
-                               (vector unsigned short)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
-  return __builtin_s390_vfaezh(__a, __b, 8);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_ne_or_0_idx(vector signed int __a, vector signed int __b) {
-  return (vector signed int)
-    __builtin_s390_vfaezf((vector unsigned int)__a,
-                          (vector unsigned int)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_or_0_idx(vector bool int __a, vector bool int __b) {
-  return __builtin_s390_vfaezf((vector unsigned int)__a,
-                               (vector unsigned int)__b, 8);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
-  return __builtin_s390_vfaezf(__a, __b, 8);
-}
-
-/*-- vec_find_any_ne_or_0_idx_cc --------------------------------------------*/
-
-static inline __ATTRS_o_ai vector signed char
-vec_find_any_ne_or_0_idx_cc(vector signed char __a, vector signed char __b,
-                            int *__cc) {
-  return (vector signed char)
-    __builtin_s390_vfaezbs((vector unsigned char)__a,
-                           (vector unsigned char)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_or_0_idx_cc(vector bool char __a, vector bool char __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezbs((vector unsigned char)__a,
-                                (vector unsigned char)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned char
-vec_find_any_ne_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezbs(__a, __b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed short
-vec_find_any_ne_or_0_idx_cc(vector signed short __a, vector signed short __b,
-                            int *__cc) {
-  return (vector signed short)
-    __builtin_s390_vfaezhs((vector unsigned short)__a,
-                           (vector unsigned short)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_or_0_idx_cc(vector bool short __a, vector bool short __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezhs((vector unsigned short)__a,
-                                (vector unsigned short)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned short
-vec_find_any_ne_or_0_idx_cc(vector unsigned short __a,
-                            vector unsigned short __b, int *__cc) {
-  return __builtin_s390_vfaezhs(__a, __b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector signed int
-vec_find_any_ne_or_0_idx_cc(vector signed int __a, vector signed int __b,
-                            int *__cc) {
-  return (vector signed int)
-    __builtin_s390_vfaezfs((vector unsigned int)__a,
-                           (vector unsigned int)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_or_0_idx_cc(vector bool int __a, vector bool int __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezfs((vector unsigned int)__a,
-                                (vector unsigned int)__b, 8, __cc);
-}
-
-static inline __ATTRS_o_ai vector unsigned int
-vec_find_any_ne_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
-                            int *__cc) {
-  return __builtin_s390_vfaezfs(__a, __b, 8, __cc);
-}
-
-#undef __constant_pow2_range
-#undef __constant_range
-#undef __constant
-#undef __ATTRS_o
-#undef __ATTRS_o_ai
-#undef __ATTRS_ai
-
-#else
-
-#error "Use -fzvector to enable vector extensions"
-
-#endif
diff --git a/linux-x86/lib64/clang/9.0.7/include/xmmintrin.h b/linux-x86/lib64/clang/9.0.7/include/xmmintrin.h
deleted file mode 100644
index 6f5517e..0000000
--- a/linux-x86/lib64/clang/9.0.7/include/xmmintrin.h
+++ /dev/null
@@ -1,3000 +0,0 @@
-/*===---- xmmintrin.h - SSE intrinsics -------------------------------------===
- *
- * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
- * See https://llvm.org/LICENSE.txt for license information.
- * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
- *
- *===-----------------------------------------------------------------------===
- */
-
-#ifndef __XMMINTRIN_H
-#define __XMMINTRIN_H
-
-#include <mmintrin.h>
-
-typedef int __v4si __attribute__((__vector_size__(16)));
-typedef float __v4sf __attribute__((__vector_size__(16)));
-typedef float __m128 __attribute__((__vector_size__(16), __aligned__(16)));
-
-typedef float __m128_u __attribute__((__vector_size__(16), __aligned__(1)));
-
-/* Unsigned types */
-typedef unsigned int __v4su __attribute__((__vector_size__(16)));
-
-/* This header should only be included in a hosted environment as it depends on
- * a standard library to provide allocation routines. */
-#if __STDC_HOSTED__
-#include <mm_malloc.h>
-#endif
-
-/* Define the default attributes for the functions in this file. */
-#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("sse"), __min_vector_width__(128)))
-#define __DEFAULT_FN_ATTRS_MMX __attribute__((__always_inline__, __nodebug__, __target__("mmx,sse"), __min_vector_width__(64)))
-
-/// Adds the 32-bit float values in the low-order bits of the operands.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VADDSS / ADDSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-///    The lower 32 bits of this operand are used in the calculation.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-///    The lower 32 bits of this operand are used in the calculation.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the sum
-///    of the lower 32 bits of both operands. The upper 96 bits are copied from
-///    the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_add_ss(__m128 __a, __m128 __b)
-{
-  __a[0] += __b[0];
-  return __a;
-}
-
-/// Adds two 128-bit vectors of [4 x float], and returns the results of
-///    the addition.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VADDPS / ADDPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \returns A 128-bit vector of [4 x float] containing the sums of both
-///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_add_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4sf)__a + (__v4sf)__b);
-}
-
-/// Subtracts the 32-bit float value in the low-order bits of the second
-///    operand from the corresponding value in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSUBSS / SUBSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the minuend. The lower 32 bits
-///    of this operand are used in the calculation.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing the subtrahend. The lower 32
-///    bits of this operand are used in the calculation.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    difference of the lower 32 bits of both operands. The upper 96 bits are
-///    copied from the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sub_ss(__m128 __a, __m128 __b)
-{
-  __a[0] -= __b[0];
-  return __a;
-}
-
-/// Subtracts each of the values of the second operand from the first
-///    operand, both of which are 128-bit vectors of [4 x float] and returns
-///    the results of the subtraction.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSUBPS / SUBPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the minuend.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing the subtrahend.
-/// \returns A 128-bit vector of [4 x float] containing the differences between
-///    both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sub_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4sf)__a - (__v4sf)__b);
-}
-
-/// Multiplies two 32-bit float values in the low-order bits of the
-///    operands.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMULSS / MULSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-///    The lower 32 bits of this operand are used in the calculation.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-///    The lower 32 bits of this operand are used in the calculation.
-/// \returns A 128-bit vector of [4 x float] containing the product of the lower
-///    32 bits of both operands. The upper 96 bits are copied from the upper 96
-///    bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_mul_ss(__m128 __a, __m128 __b)
-{
-  __a[0] *= __b[0];
-  return __a;
-}
-
-/// Multiplies two 128-bit vectors of [4 x float] and returns the
-///    results of the multiplication.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMULPS / MULPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \returns A 128-bit vector of [4 x float] containing the products of both
-///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_mul_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4sf)__a * (__v4sf)__b);
-}
-
-/// Divides the value in the low-order 32 bits of the first operand by
-///    the corresponding value in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VDIVSS / DIVSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the dividend. The lower 32
-///    bits of this operand are used in the calculation.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing the divisor. The lower 32 bits
-///    of this operand are used in the calculation.
-/// \returns A 128-bit vector of [4 x float] containing the quotients of the
-///    lower 32 bits of both operands. The upper 96 bits are copied from the
-///    upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_div_ss(__m128 __a, __m128 __b)
-{
-  __a[0] /= __b[0];
-  return __a;
-}
-
-/// Divides two 128-bit vectors of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VDIVPS / DIVPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the dividend.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing the divisor.
-/// \returns A 128-bit vector of [4 x float] containing the quotients of both
-///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_div_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4sf)__a / (__v4sf)__b);
-}
-
-/// Calculates the square root of the value stored in the low-order bits
-///    of a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSQRTSS / SQRTSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the calculation.
-/// \returns A 128-bit vector of [4 x float] containing the square root of the
-///    value in the low-order bits of the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sqrt_ss(__m128 __a)
-{
-  return (__m128)__builtin_ia32_sqrtss((__v4sf)__a);
-}
-
-/// Calculates the square roots of the values stored in a 128-bit vector
-///    of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSQRTPS / SQRTPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the square roots of the
-///    values in the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sqrt_ps(__m128 __a)
-{
-  return __builtin_ia32_sqrtps((__v4sf)__a);
-}
-
-/// Calculates the approximate reciprocal of the value stored in the
-///    low-order bits of a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VRCPSS / RCPSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the calculation.
-/// \returns A 128-bit vector of [4 x float] containing the approximate
-///    reciprocal of the value in the low-order bits of the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_rcp_ss(__m128 __a)
-{
-  return (__m128)__builtin_ia32_rcpss((__v4sf)__a);
-}
-
-/// Calculates the approximate reciprocals of the values stored in a
-///    128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VRCPPS / RCPPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the approximate
-///    reciprocals of the values in the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_rcp_ps(__m128 __a)
-{
-  return (__m128)__builtin_ia32_rcpps((__v4sf)__a);
-}
-
-/// Calculates the approximate reciprocal of the square root of the value
-///    stored in the low-order bits of a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VRSQRTSS / RSQRTSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the calculation.
-/// \returns A 128-bit vector of [4 x float] containing the approximate
-///    reciprocal of the square root of the value in the low-order bits of the
-///    operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_rsqrt_ss(__m128 __a)
-{
-  return __builtin_ia32_rsqrtss((__v4sf)__a);
-}
-
-/// Calculates the approximate reciprocals of the square roots of the
-///    values stored in a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VRSQRTPS / RSQRTPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the approximate
-///    reciprocals of the square roots of the values in the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_rsqrt_ps(__m128 __a)
-{
-  return __builtin_ia32_rsqrtps((__v4sf)__a);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands and returns the lesser value in the low-order bits of the
-///    vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMINSS / MINSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    minimum value between both operands. The upper 96 bits are copied from
-///    the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_min_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_minss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 128-bit vectors of [4 x float] and returns the lesser
-///    of each pair of values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMINPS / MINPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands.
-/// \returns A 128-bit vector of [4 x float] containing the minimum values
-///    between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_min_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_minps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands and returns the greater value in the low-order bits of a 128-bit
-///    vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMAXSS / MAXSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    maximum value between both operands. The upper 96 bits are copied from
-///    the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_max_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_maxss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 128-bit vectors of [4 x float] and returns the greater
-///    of each pair of values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMAXPS / MAXPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands.
-/// \returns A 128-bit vector of [4 x float] containing the maximum values
-///    between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_max_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_maxps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit vectors of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VANDPS / ANDPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector containing one of the source operands.
-/// \param __b
-///    A 128-bit vector containing one of the source operands.
-/// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
-///    values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_and_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4su)__a & (__v4su)__b);
-}
-
-/// Performs a bitwise AND of two 128-bit vectors of [4 x float], using
-///    the one's complement of the values contained in the first source
-///    operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VANDNPS / ANDNPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the first source operand. The
-///    one's complement of this value is used in the bitwise AND.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing the second source operand.
-/// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
-///    one's complement of the first operand and the values in the second
-///    operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_andnot_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)(~(__v4su)__a & (__v4su)__b);
-}
-
-/// Performs a bitwise OR of two 128-bit vectors of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VORPS / ORPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \returns A 128-bit vector of [4 x float] containing the bitwise OR of the
-///    values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_or_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4su)__a | (__v4su)__b);
-}
-
-/// Performs a bitwise exclusive OR of two 128-bit vectors of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the source operands.
-/// \returns A 128-bit vector of [4 x float] containing the bitwise exclusive OR
-///    of the values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_xor_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)((__v4su)__a ^ (__v4su)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands for equality and returns the result of the comparison in the
-///    low-order bits of a vector [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPEQSS / CMPEQSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpeq_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpeqss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] for equality.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPEQPS / CMPEQPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpeq_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpeqps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is less than the
-///    corresponding value in the second operand and returns the result of the
-///    comparison in the low-order bits of a vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTSS / CMPLTSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmplt_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpltss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are less than those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTPS / CMPLTPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmplt_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpltps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is less than or
-///    equal to the corresponding value in the second operand and returns the
-///    result of the comparison in the low-order bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLESS / CMPLESS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmple_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpless((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are less than or equal to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLEPS / CMPLEPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmple_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpleps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is greater than
-///    the corresponding value in the second operand and returns the result of
-///    the comparison in the low-order bits of a vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTSS / CMPLTSS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpgt_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_shufflevector((__v4sf)__a,
-                                         (__v4sf)__builtin_ia32_cmpltss((__v4sf)__b, (__v4sf)__a),
-                                         4, 1, 2, 3);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are greater than those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLTPS / CMPLTPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpgt_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpltps((__v4sf)__b, (__v4sf)__a);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is greater than
-///    or equal to the corresponding value in the second operand and returns
-///    the result of the comparison in the low-order bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLESS / CMPLESS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpge_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_shufflevector((__v4sf)__a,
-                                         (__v4sf)__builtin_ia32_cmpless((__v4sf)__b, (__v4sf)__a),
-                                         4, 1, 2, 3);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are greater than or equal to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPLEPS / CMPLEPS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpge_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpleps((__v4sf)__b, (__v4sf)__a);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands for inequality and returns the result of the comparison in the
-///    low-order bits of a vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNEQSS / CMPNEQSS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpneq_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpneqss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] for inequality.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNEQPS / CMPNEQPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpneq_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpneqps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is not less than
-///    the corresponding value in the second operand and returns the result of
-///    the comparison in the low-order bits of a vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTSS / CMPNLTSS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnlt_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnltss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are not less than those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTPS / CMPNLTPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnlt_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnltps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is not less than
-///    or equal to the corresponding value in the second operand and returns
-///    the result of the comparison in the low-order bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLESS / CMPNLESS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnle_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnless((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are not less than or equal to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLEPS / CMPNLEPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnle_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnleps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is not greater
-///    than the corresponding value in the second operand and returns the
-///    result of the comparison in the low-order bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTSS / CMPNLTSS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpngt_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_shufflevector((__v4sf)__a,
-                                         (__v4sf)__builtin_ia32_cmpnltss((__v4sf)__b, (__v4sf)__a),
-                                         4, 1, 2, 3);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are not greater than those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLTPS / CMPNLTPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpngt_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnltps((__v4sf)__b, (__v4sf)__a);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is not greater
-///    than or equal to the corresponding value in the second operand and
-///    returns the result of the comparison in the low-order bits of a vector
-///    of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLESS / CMPNLESS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnge_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_shufflevector((__v4sf)__a,
-                                         (__v4sf)__builtin_ia32_cmpnless((__v4sf)__b, (__v4sf)__a),
-                                         4, 1, 2, 3);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are not greater than or equal to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPNLEPS / CMPNLEPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpnge_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpnleps((__v4sf)__b, (__v4sf)__a);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is ordered with
-///    respect to the corresponding value in the second operand and returns the
-///    result of the comparison in the low-order bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPORDSS / CMPORDSS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpord_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpordss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are ordered with respect to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPORDPS / CMPORDPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpord_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpordps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the value in the first operand is unordered
-///    with respect to the corresponding value in the second operand and
-///    returns the result of the comparison in the low-order bits of a vector
-///    of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPUNORDSS / CMPUNORDSS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float] containing one of the operands. The lower
-///    32 bits of this operand are used in the comparison.
-/// \returns A 128-bit vector of [4 x float] containing the comparison results
-///    in the low-order bits.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpunord_ss(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpunordss((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares each of the corresponding 32-bit float values of the
-///    128-bit vectors of [4 x float] to determine if the values in the first
-///    operand are unordered with respect to those in the second operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCMPUNORDPS / CMPUNORDPS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 128-bit vector of [4 x float].
-/// \returns A 128-bit vector of [4 x float] containing the comparison results.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cmpunord_ps(__m128 __a, __m128 __b)
-{
-  return (__m128)__builtin_ia32_cmpunordps((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands for equality and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the
-///    two lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comieq_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comieq((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the first operand is less than the second
-///    operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comilt_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comilt((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the first operand is less than or equal to the
-///    second operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comile_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comile((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the first operand is greater than the second
-///    operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the
-///     two lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comigt_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comigt((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the first operand is greater than or equal to
-///    the second operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comige_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comige((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Compares two 32-bit float values in the low-order bits of both
-///    operands to determine if the first operand is not equal to the second
-///    operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 1 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCOMISS / COMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the
-///     two lower 32-bit values is NaN, 1 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_comineq_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_comineq((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine equality and returns
-///    the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomieq_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomieq((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine if the first operand is
-///    less than the second operand and returns the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomilt_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomilt((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine if the first operand is
-///    less than or equal to the second operand and returns the result of the
-///    comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomile_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomile((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine if the first operand is
-///    greater than the second operand and returns the result of the
-///    comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomigt_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomigt((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine if the first operand is
-///    greater than or equal to the second operand and returns the result of
-///    the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 0 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///     lower 32-bit values is NaN, 0 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomige_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomige((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Performs an unordered comparison of two 32-bit float values using
-///    the low-order bits of both operands to determine inequality and returns
-///    the result of the comparison.
-///
-///    If either of the two lower 32-bit values is NaN, 1 is returned.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUCOMISS / UCOMISS </c> instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \param __b
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the comparison.
-/// \returns An integer containing the comparison results. If either of the two
-///    lower 32-bit values is NaN, 1 is returned.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_ucomineq_ss(__m128 __a, __m128 __b)
-{
-  return __builtin_ia32_ucomineq((__v4sf)__a, (__v4sf)__b);
-}
-
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 32-bit integer.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 32-bit integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvtss_si32(__m128 __a)
-{
-  return __builtin_ia32_cvtss2si((__v4sf)__a);
-}
-
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 32-bit integer.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 32-bit integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvt_ss2si(__m128 __a)
-{
-  return _mm_cvtss_si32(__a);
-}
-
-#ifdef __x86_64__
-
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 64-bit integer.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSS2SI / CVTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 64-bit integer containing the converted value.
-static __inline__ long long __DEFAULT_FN_ATTRS
-_mm_cvtss_si64(__m128 __a)
-{
-  return __builtin_ia32_cvtss2si64((__v4sf)__a);
-}
-
-#endif
-
-/// Converts two low-order float values in a 128-bit vector of
-///    [4 x float] into a 64-bit vector of [2 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPS2PI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtps_pi32(__m128 __a)
-{
-  return (__m64)__builtin_ia32_cvtps2pi((__v4sf)__a);
-}
-
-/// Converts two low-order float values in a 128-bit vector of
-///    [4 x float] into a 64-bit vector of [2 x i32].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPS2PI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvt_ps2pi(__m128 __a)
-{
-  return _mm_cvtps_pi32(__a);
-}
-
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 32-bit integer, truncating the result when it is
-///    inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 32-bit integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvttss_si32(__m128 __a)
-{
-  return __builtin_ia32_cvttss2si((__v4sf)__a);
-}
-
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 32-bit integer, truncating the result when it is
-///    inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 32-bit integer containing the converted value.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_cvtt_ss2si(__m128 __a)
-{
-  return _mm_cvttss_si32(__a);
-}
-
-#ifdef __x86_64__
-/// Converts a float value contained in the lower 32 bits of a vector of
-///    [4 x float] into a 64-bit integer, truncating the result when it is
-///    inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTTSS2SI / CVTTSS2SI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the conversion.
-/// \returns A 64-bit integer containing the converted value.
-static __inline__ long long __DEFAULT_FN_ATTRS
-_mm_cvttss_si64(__m128 __a)
-{
-  return __builtin_ia32_cvttss2si64((__v4sf)__a);
-}
-#endif
-
-/// Converts two low-order float values in a 128-bit vector of
-///    [4 x float] into a 64-bit vector of [2 x i32], truncating the result
-///    when it is inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTTPS2PI / VTTPS2PI </c>
-///   instructions.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvttps_pi32(__m128 __a)
-{
-  return (__m64)__builtin_ia32_cvttps2pi((__v4sf)__a);
-}
-
-/// Converts two low-order float values in a 128-bit vector of [4 x
-///    float] into a 64-bit vector of [2 x i32], truncating the result when it
-///    is inexact.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTTPS2PI </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtt_ps2pi(__m128 __a)
-{
-  return _mm_cvttps_pi32(__a);
-}
-
-/// Converts a 32-bit signed integer value into a floating point value
-///    and writes it to the lower 32 bits of the destination. The remaining
-///    higher order elements of the destination vector are copied from the
-///    corresponding elements in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 32-bit signed integer operand containing the value to be converted.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    converted value of the second operand. The upper 96 bits are copied from
-///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtsi32_ss(__m128 __a, int __b)
-{
-  __a[0] = __b;
-  return __a;
-}
-
-/// Converts a 32-bit signed integer value into a floating point value
-///    and writes it to the lower 32 bits of the destination. The remaining
-///    higher order elements of the destination are copied from the
-///    corresponding elements in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 32-bit signed integer operand containing the value to be converted.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    converted value of the second operand. The upper 96 bits are copied from
-///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvt_si2ss(__m128 __a, int __b)
-{
-  return _mm_cvtsi32_ss(__a, __b);
-}
-
-#ifdef __x86_64__
-
-/// Converts a 64-bit signed integer value into a floating point value
-///    and writes it to the lower 32 bits of the destination. The remaining
-///    higher order elements of the destination are copied from the
-///    corresponding elements in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VCVTSI2SS / CVTSI2SS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 64-bit signed integer operand containing the value to be converted.
-/// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
-///    converted value of the second operand. The upper 96 bits are copied from
-///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtsi64_ss(__m128 __a, long long __b)
-{
-  __a[0] = __b;
-  return __a;
-}
-
-#endif
-
-/// Converts two elements of a 64-bit vector of [2 x i32] into two
-///    floating point values and writes them to the lower 64-bits of the
-///    destination. The remaining higher order elements of the destination are
-///    copied from the corresponding elements in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 64-bit vector of [2 x i32]. The elements in this vector are converted
-///    and written to the corresponding low-order elements in the destination.
-/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
-///    converted value of the second operand. The upper 64 bits are copied from
-///    the upper 64 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpi32_ps(__m128 __a, __m64 __b)
-{
-  return __builtin_ia32_cvtpi2ps((__v4sf)__a, (__v2si)__b);
-}
-
-/// Converts two elements of a 64-bit vector of [2 x i32] into two
-///    floating point values and writes them to the lower 64-bits of the
-///    destination. The remaining higher order elements of the destination are
-///    copied from the corresponding elements in the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float].
-/// \param __b
-///    A 64-bit vector of [2 x i32]. The elements in this vector are converted
-///    and written to the corresponding low-order elements in the destination.
-/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
-///    converted value from the second operand. The upper 64 bits are copied
-///    from the upper 64 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvt_pi2ps(__m128 __a, __m64 __b)
-{
-  return _mm_cvtpi32_ps(__a, __b);
-}
-
-/// Extracts a float value contained in the lower 32 bits of a vector of
-///    [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
-///    used in the extraction.
-/// \returns A 32-bit float containing the extracted value.
-static __inline__ float __DEFAULT_FN_ATTRS
-_mm_cvtss_f32(__m128 __a)
-{
-  return __a[0];
-}
-
-/// Loads two packed float values from the address \a __p into the
-///     high-order bits of a 128-bit vector of [4 x float]. The low-order bits
-///     are copied from the low-order bits of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVHPD / MOVHPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. Bits [63:0] are written to bits [63:0]
-///    of the destination.
-/// \param __p
-///    A pointer to two packed float values. Bits [63:0] are written to bits
-///    [127:64] of the destination.
-/// \returns A 128-bit vector of [4 x float] containing the moved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_loadh_pi(__m128 __a, const __m64 *__p)
-{
-  typedef float __mm_loadh_pi_v2f32 __attribute__((__vector_size__(8)));
-  struct __mm_loadh_pi_struct {
-    __mm_loadh_pi_v2f32 __u;
-  } __attribute__((__packed__, __may_alias__));
-  __mm_loadh_pi_v2f32 __b = ((struct __mm_loadh_pi_struct*)__p)->__u;
-  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
-  return __builtin_shufflevector(__a, __bb, 0, 1, 4, 5);
-}
-
-/// Loads two packed float values from the address \a __p into the
-///    low-order bits of a 128-bit vector of [4 x float]. The high-order bits
-///    are copied from the high-order bits of the first operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVLPD / MOVLPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. Bits [127:64] are written to bits
-///    [127:64] of the destination.
-/// \param __p
-///    A pointer to two packed float values. Bits [63:0] are written to bits
-///    [63:0] of the destination.
-/// \returns A 128-bit vector of [4 x float] containing the moved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_loadl_pi(__m128 __a, const __m64 *__p)
-{
-  typedef float __mm_loadl_pi_v2f32 __attribute__((__vector_size__(8)));
-  struct __mm_loadl_pi_struct {
-    __mm_loadl_pi_v2f32 __u;
-  } __attribute__((__packed__, __may_alias__));
-  __mm_loadl_pi_v2f32 __b = ((struct __mm_loadl_pi_struct*)__p)->__u;
-  __m128 __bb = __builtin_shufflevector(__b, __b, 0, 1, 0, 1);
-  return __builtin_shufflevector(__a, __bb, 4, 5, 2, 3);
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
-///    32 bits of the vector are initialized with the single-precision
-///    floating-point value loaded from a specified memory location. The upper
-///    96 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 32-bit memory location containing a single-precision
-///    floating-point value.
-/// \returns An initialized 128-bit floating-point vector of [4 x float]. The
-///    lower 32 bits contain the value loaded from the memory location. The
-///    upper 96 bits are set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_load_ss(const float *__p)
-{
-  struct __mm_load_ss_struct {
-    float __u;
-  } __attribute__((__packed__, __may_alias__));
-  float __u = ((struct __mm_load_ss_struct*)__p)->__u;
-  return __extension__ (__m128){ __u, 0, 0, 0 };
-}
-
-/// Loads a 32-bit float value and duplicates it to all four vector
-///    elements of a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VBROADCASTSS / MOVSS + shuffling </c>
-///    instruction.
-///
-/// \param __p
-///    A pointer to a float value to be loaded and duplicated.
-/// \returns A 128-bit vector of [4 x float] containing the loaded and
-///    duplicated values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_load1_ps(const float *__p)
-{
-  struct __mm_load1_ps_struct {
-    float __u;
-  } __attribute__((__packed__, __may_alias__));
-  float __u = ((struct __mm_load1_ps_struct*)__p)->__u;
-  return __extension__ (__m128){ __u, __u, __u, __u };
-}
-
-#define        _mm_load_ps1(p) _mm_load1_ps(p)
-
-/// Loads a 128-bit floating-point vector of [4 x float] from an aligned
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location has to be 128-bit aligned.
-/// \returns A 128-bit vector of [4 x float] containing the loaded values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_load_ps(const float *__p)
-{
-  return *(__m128*)__p;
-}
-
-/// Loads a 128-bit floating-point vector of [4 x float] from an
-///    unaligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVUPS / MOVUPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \returns A 128-bit vector of [4 x float] containing the loaded values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_loadu_ps(const float *__p)
-{
-  struct __loadu_ps {
-    __m128_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  return ((struct __loadu_ps*)__p)->__v;
-}
-
-/// Loads four packed float values, in reverse order, from an aligned
-///    memory location to 32-bit elements in a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS + shuffling </c>
-///    instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location has to be 128-bit aligned.
-/// \returns A 128-bit vector of [4 x float] containing the moved values, loaded
-///    in reverse order.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_loadr_ps(const float *__p)
-{
-  __m128 __a = _mm_load_ps(__p);
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 3, 2, 1, 0);
-}
-
-/// Create a 128-bit vector of [4 x float] with undefined values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic has no corresponding instruction.
-///
-/// \returns A 128-bit vector of [4 x float] containing undefined values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_undefined_ps(void)
-{
-  return (__m128)__builtin_ia32_undef128();
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
-///    32 bits of the vector are initialized with the specified single-precision
-///    floating-point value. The upper 96 bits are set to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
-///
-/// \param __w
-///    A single-precision floating-point value used to initialize the lower 32
-///    bits of the result.
-/// \returns An initialized 128-bit floating-point vector of [4 x float]. The
-///    lower 32 bits contain the value provided in the source operand. The
-///    upper 96 bits are set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ss(float __w)
-{
-  return __extension__ (__m128){ __w, 0, 0, 0 };
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float], with each
-///    of the four single-precision floating-point vector elements set to the
-///    specified single-precision floating-point value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPERMILPS / PERMILPS </c> instruction.
-///
-/// \param __w
-///    A single-precision floating-point value used to initialize each vector
-///    element of the result.
-/// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set1_ps(float __w)
-{
-  return __extension__ (__m128){ __w, __w, __w, __w };
-}
-
-/* Microsoft specific. */
-/// Constructs a 128-bit floating-point vector of [4 x float], with each
-///    of the four single-precision floating-point vector elements set to the
-///    specified single-precision floating-point value.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPERMILPS / PERMILPS </c> instruction.
-///
-/// \param __w
-///    A single-precision floating-point value used to initialize each vector
-///    element of the result.
-/// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ps1(float __w)
-{
-    return _mm_set1_ps(__w);
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]
-///    initialized with the specified single-precision floating-point values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __z
-///    A single-precision floating-point value used to initialize bits [127:96]
-///    of the result.
-/// \param __y
-///    A single-precision floating-point value used to initialize bits [95:64]
-///    of the result.
-/// \param __x
-///    A single-precision floating-point value used to initialize bits [63:32]
-///    of the result.
-/// \param __w
-///    A single-precision floating-point value used to initialize bits [31:0]
-///    of the result.
-/// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ps(float __z, float __y, float __x, float __w)
-{
-  return __extension__ (__m128){ __w, __x, __y, __z };
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float],
-///    initialized in reverse order with the specified 32-bit single-precision
-///    float-point values.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic is a utility function and does not correspond to a specific
-///    instruction.
-///
-/// \param __z
-///    A single-precision floating-point value used to initialize bits [31:0]
-///    of the result.
-/// \param __y
-///    A single-precision floating-point value used to initialize bits [63:32]
-///    of the result.
-/// \param __x
-///    A single-precision floating-point value used to initialize bits [95:64]
-///    of the result.
-/// \param __w
-///    A single-precision floating-point value used to initialize bits [127:96]
-///    of the result.
-/// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_setr_ps(float __z, float __y, float __x, float __w)
-{
-  return __extension__ (__m128){ __z, __y, __x, __w };
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float] initialized
-///    to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VXORPS / XORPS </c> instruction.
-///
-/// \returns An initialized 128-bit floating-point vector of [4 x float] with
-///    all elements set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_setzero_ps(void)
-{
-  return __extension__ (__m128){ 0, 0, 0, 0 };
-}
-
-/// Stores the upper 64 bits of a 128-bit vector of [4 x float] to a
-///    memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VPEXTRQ / PEXTRQ </c> instruction.
-///
-/// \param __p
-///    A pointer to a 64-bit memory location.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeh_pi(__m64 *__p, __m128 __a)
-{
-  __builtin_ia32_storehps((__v2si *)__p, (__v4sf)__a);
-}
-
-/// Stores the lower 64 bits of a 128-bit vector of [4 x float] to a
-///     memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVLPS / MOVLPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a memory location that will receive the float values.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storel_pi(__m64 *__p, __m128 __a)
-{
-  __builtin_ia32_storelps((__v2si *)__p, (__v4sf)__a);
-}
-
-/// Stores the lower 32 bits of a 128-bit vector of [4 x float] to a
-///     memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVSS / MOVSS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 32-bit memory location.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_ss(float *__p, __m128 __a)
-{
-  struct __mm_store_ss_struct {
-    float __u;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __mm_store_ss_struct*)__p)->__u = __a[0];
-}
-
-/// Stores a 128-bit vector of [4 x float] to an unaligned memory
-///    location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVUPS / MOVUPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location does not have to be aligned.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storeu_ps(float *__p, __m128 __a)
-{
-  struct __storeu_ps {
-    __m128_u __v;
-  } __attribute__((__packed__, __may_alias__));
-  ((struct __storeu_ps*)__p)->__v = __a;
-}
-
-/// Stores a 128-bit vector of [4 x float] into an aligned memory
-///    location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location has to be 16-byte aligned.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_ps(float *__p, __m128 __a)
-{
-  *(__m128*)__p = __a;
-}
-
-/// Stores the lower 32 bits of a 128-bit vector of [4 x float] into
-///    four contiguous elements in an aligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to <c> VMOVAPS / MOVAPS + shuffling </c>
-///    instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location.
-/// \param __a
-///    A 128-bit vector of [4 x float] whose lower 32 bits are stored to each
-///    of the four contiguous elements pointed by \a __p.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store1_ps(float *__p, __m128 __a)
-{
-  __a = __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 0, 0, 0, 0);
-  _mm_store_ps(__p, __a);
-}
-
-/// Stores the lower 32 bits of a 128-bit vector of [4 x float] into
-///    four contiguous elements in an aligned memory location.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to <c> VMOVAPS / MOVAPS + shuffling </c>
-///    instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location.
-/// \param __a
-///    A 128-bit vector of [4 x float] whose lower 32 bits are stored to each
-///    of the four contiguous elements pointed by \a __p.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_store_ps1(float *__p, __m128 __a)
-{
-  _mm_store1_ps(__p, __a);
-}
-
-/// Stores float values from a 128-bit vector of [4 x float] to an
-///    aligned memory location in reverse order.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVAPS / MOVAPS + shuffling </c>
-///    instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit memory location. The address of the memory
-///    location has to be 128-bit aligned.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_storer_ps(float *__p, __m128 __a)
-{
-  __a = __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 3, 2, 1, 0);
-  _mm_store_ps(__p, __a);
-}
-
-#define _MM_HINT_ET0 7
-#define _MM_HINT_ET1 6
-#define _MM_HINT_T0  3
-#define _MM_HINT_T1  2
-#define _MM_HINT_T2  1
-#define _MM_HINT_NTA 0
-
-#ifndef _MSC_VER
-/* FIXME: We have to #define this because "sel" must be a constant integer, and
-   Sema doesn't do any form of constant propagation yet. */
-
-/// Loads one cache line of data from the specified address to a location
-///    closer to the processor.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// void _mm_prefetch(const void * a, const int sel);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> PREFETCHNTA </c> instruction.
-///
-/// \param a
-///    A pointer to a memory location containing a cache line of data.
-/// \param sel
-///    A predefined integer constant specifying the type of prefetch
-///    operation: \n
-///    _MM_HINT_NTA: Move data using the non-temporal access (NTA) hint. The
-///    PREFETCHNTA instruction will be generated. \n
-///    _MM_HINT_T0: Move data using the T0 hint. The PREFETCHT0 instruction will
-///    be generated. \n
-///    _MM_HINT_T1: Move data using the T1 hint. The PREFETCHT1 instruction will
-///    be generated. \n
-///    _MM_HINT_T2: Move data using the T2 hint. The PREFETCHT2 instruction will
-///    be generated.
-#define _mm_prefetch(a, sel) (__builtin_prefetch((void *)(a), \
-                                                 ((sel) >> 2) & 1, (sel) & 0x3))
-#endif
-
-/// Stores a 64-bit integer in the specified aligned memory location. To
-///    minimize caching, the data is flagged as non-temporal (unlikely to be
-///    used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MOVNTQ </c> instruction.
-///
-/// \param __p
-///    A pointer to an aligned memory location used to store the register value.
-/// \param __a
-///    A 64-bit integer containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS_MMX
-_mm_stream_pi(__m64 *__p, __m64 __a)
-{
-  __builtin_ia32_movntq(__p, __a);
-}
-
-/// Moves packed float values from a 128-bit vector of [4 x float] to a
-///    128-bit aligned memory location. To minimize caching, the data is flagged
-///    as non-temporal (unlikely to be used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVNTPS / MOVNTPS </c> instruction.
-///
-/// \param __p
-///    A pointer to a 128-bit aligned memory location that will receive the
-///    single-precision floating-point values.
-/// \param __a
-///    A 128-bit vector of [4 x float] containing the values to be moved.
-static __inline__ void __DEFAULT_FN_ATTRS
-_mm_stream_ps(float *__p, __m128 __a)
-{
-  __builtin_nontemporal_store((__v4sf)__a, (__v4sf*)__p);
-}
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-/// Forces strong memory ordering (serialization) between store
-///    instructions preceding this instruction and store instructions following
-///    this instruction, ensuring the system completes all previous stores
-///    before executing subsequent stores.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> SFENCE </c> instruction.
-///
-void _mm_sfence(void);
-
-#if defined(__cplusplus)
-} // extern "C"
-#endif
-
-/// Extracts 16-bit element from a 64-bit vector of [4 x i16] and
-///    returns it, as specified by the immediate integer operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// int _mm_extract_pi16(__m64 a, int n);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VPEXTRW / PEXTRW </c> instruction.
-///
-/// \param a
-///    A 64-bit vector of [4 x i16].
-/// \param n
-///    An immediate integer operand that determines which bits are extracted: \n
-///    0: Bits [15:0] are copied to the destination. \n
-///    1: Bits [31:16] are copied to the destination. \n
-///    2: Bits [47:32] are copied to the destination. \n
-///    3: Bits [63:48] are copied to the destination.
-/// \returns A 16-bit integer containing the extracted 16 bits of packed data.
-#define _mm_extract_pi16(a, n) \
-  (int)__builtin_ia32_vec_ext_v4hi((__m64)a, (int)n)
-
-/// Copies data from the 64-bit vector of [4 x i16] to the destination,
-///    and inserts the lower 16-bits of an integer operand at the 16-bit offset
-///    specified by the immediate operand \a n.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m64 _mm_insert_pi16(__m64 a, int d, int n);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> PINSRW </c> instruction.
-///
-/// \param a
-///    A 64-bit vector of [4 x i16].
-/// \param d
-///    An integer. The lower 16-bit value from this operand is written to the
-///    destination at the offset specified by operand \a n.
-/// \param n
-///    An immediate integer operant that determines which the bits to be used
-///    in the destination. \n
-///    0: Bits [15:0] are copied to the destination. \n
-///    1: Bits [31:16] are copied to the destination. \n
-///    2: Bits [47:32] are copied to the destination. \n
-///    3: Bits [63:48] are copied to the destination.  \n
-///    The remaining bits in the destination are copied from the corresponding
-///    bits in operand \a a.
-/// \returns A 64-bit integer vector containing the copied packed data from the
-///    operands.
-#define _mm_insert_pi16(a, d, n) \
-  (__m64)__builtin_ia32_vec_set_v4hi((__m64)a, (int)d, (int)n)
-
-/// Compares each of the corresponding packed 16-bit integer values of
-///    the 64-bit integer vectors, and writes the greater value to the
-///    corresponding bits in the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMAXSW </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_max_pi16(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pmaxsw((__v4hi)__a, (__v4hi)__b);
-}
-
-/// Compares each of the corresponding packed 8-bit unsigned integer
-///    values of the 64-bit integer vectors, and writes the greater value to the
-///    corresponding bits in the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMAXUB </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_max_pu8(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pmaxub((__v8qi)__a, (__v8qi)__b);
-}
-
-/// Compares each of the corresponding packed 16-bit integer values of
-///    the 64-bit integer vectors, and writes the lesser value to the
-///    corresponding bits in the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMINSW </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_min_pi16(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pminsw((__v4hi)__a, (__v4hi)__b);
-}
-
-/// Compares each of the corresponding packed 8-bit unsigned integer
-///    values of the 64-bit integer vectors, and writes the lesser value to the
-///    corresponding bits in the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMINUB </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_min_pu8(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pminub((__v8qi)__a, (__v8qi)__b);
-}
-
-/// Takes the most significant bit from each 8-bit element in a 64-bit
-///    integer vector to create an 8-bit mask value. Zero-extends the value to
-///    32-bit integer and writes it to the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMOVMSKB </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing the values with bits to be extracted.
-/// \returns The most significant bit from each 8-bit element in \a __a,
-///    written to bits [7:0].
-static __inline__ int __DEFAULT_FN_ATTRS_MMX
-_mm_movemask_pi8(__m64 __a)
-{
-  return __builtin_ia32_pmovmskb((__v8qi)__a);
-}
-
-/// Multiplies packed 16-bit unsigned integer values and writes the
-///    high-order 16 bits of each 32-bit product to the corresponding bits in
-///    the destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PMULHUW </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the products of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_mulhi_pu16(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pmulhuw((__v4hi)__a, (__v4hi)__b);
-}
-
-/// Shuffles the 4 16-bit integers from a 64-bit integer vector to the
-///    destination, as specified by the immediate value operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m64 _mm_shuffle_pi16(__m64 a, const int n);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> PSHUFW </c> instruction.
-///
-/// \param a
-///    A 64-bit integer vector containing the values to be shuffled.
-/// \param n
-///    An immediate value containing an 8-bit value specifying which elements to
-///    copy from \a a. The destinations within the 64-bit destination are
-///    assigned values as follows: \n
-///    Bits [1:0] are used to assign values to bits [15:0] in the
-///    destination. \n
-///    Bits [3:2] are used to assign values to bits [31:16] in the
-///    destination. \n
-///    Bits [5:4] are used to assign values to bits [47:32] in the
-///    destination. \n
-///    Bits [7:6] are used to assign values to bits [63:48] in the
-///    destination. \n
-///    Bit value assignments: \n
-///    00: assigned from bits [15:0] of \a a. \n
-///    01: assigned from bits [31:16] of \a a. \n
-///    10: assigned from bits [47:32] of \a a. \n
-///    11: assigned from bits [63:48] of \a a.
-/// \returns A 64-bit integer vector containing the shuffled values.
-#define _mm_shuffle_pi16(a, n) \
-  (__m64)__builtin_ia32_pshufw((__v4hi)(__m64)(a), (n))
-
-/// Conditionally copies the values from each 8-bit element in the first
-///    64-bit integer vector operand to the specified memory location, as
-///    specified by the most significant bit in the corresponding element in the
-///    second 64-bit integer vector operand.
-///
-///    To minimize caching, the data is flagged as non-temporal
-///    (unlikely to be used again soon).
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> MASKMOVQ </c> instruction.
-///
-/// \param __d
-///    A 64-bit integer vector containing the values with elements to be copied.
-/// \param __n
-///    A 64-bit integer vector operand. The most significant bit from each 8-bit
-///    element determines whether the corresponding element in operand \a __d
-///    is copied. If the most significant bit of a given element is 1, the
-///    corresponding element in operand \a __d is copied.
-/// \param __p
-///    A pointer to a 64-bit memory location that will receive the conditionally
-///    copied integer values. The address of the memory location does not have
-///    to be aligned.
-static __inline__ void __DEFAULT_FN_ATTRS_MMX
-_mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
-{
-  __builtin_ia32_maskmovq((__v8qi)__d, (__v8qi)__n, __p);
-}
-
-/// Computes the rounded averages of the packed unsigned 8-bit integer
-///    values and writes the averages to the corresponding bits in the
-///    destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PAVGB </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the averages of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_avg_pu8(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pavgb((__v8qi)__a, (__v8qi)__b);
-}
-
-/// Computes the rounded averages of the packed unsigned 16-bit integer
-///    values and writes the averages to the corresponding bits in the
-///    destination.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PAVGW </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector containing the averages of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_avg_pu16(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_pavgw((__v4hi)__a, (__v4hi)__b);
-}
-
-/// Subtracts the corresponding 8-bit unsigned integer values of the two
-///    64-bit vector operands and computes the absolute value for each of the
-///    difference. Then sum of the 8 absolute differences is written to the
-///    bits [15:0] of the destination; the remaining bits [63:16] are cleared.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> PSADBW </c> instruction.
-///
-/// \param __a
-///    A 64-bit integer vector containing one of the source operands.
-/// \param __b
-///    A 64-bit integer vector containing one of the source operands.
-/// \returns A 64-bit integer vector whose lower 16 bits contain the sums of the
-///    sets of absolute differences between both operands. The upper bits are
-///    cleared.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_sad_pu8(__m64 __a, __m64 __b)
-{
-  return (__m64)__builtin_ia32_psadbw((__v8qi)__a, (__v8qi)__b);
-}
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-/// Returns the contents of the MXCSR register as a 32-bit unsigned
-///    integer value.
-///
-///    There are several groups of macros associated with this
-///    intrinsic, including:
-///    <ul>
-///    <li>
-///      For checking exception states: _MM_EXCEPT_INVALID, _MM_EXCEPT_DIV_ZERO,
-///      _MM_EXCEPT_DENORM, _MM_EXCEPT_OVERFLOW, _MM_EXCEPT_UNDERFLOW,
-///      _MM_EXCEPT_INEXACT. There is a convenience wrapper
-///      _MM_GET_EXCEPTION_STATE().
-///    </li>
-///    <li>
-///      For checking exception masks: _MM_MASK_UNDERFLOW, _MM_MASK_OVERFLOW,
-///      _MM_MASK_INVALID, _MM_MASK_DENORM, _MM_MASK_DIV_ZERO, _MM_MASK_INEXACT.
-///      There is a convenience wrapper _MM_GET_EXCEPTION_MASK().
-///    </li>
-///    <li>
-///      For checking rounding modes: _MM_ROUND_NEAREST, _MM_ROUND_DOWN,
-///      _MM_ROUND_UP, _MM_ROUND_TOWARD_ZERO. There is a convenience wrapper
-///      _MM_GET_ROUNDING_MODE().
-///    </li>
-///    <li>
-///      For checking flush-to-zero mode: _MM_FLUSH_ZERO_ON, _MM_FLUSH_ZERO_OFF.
-///      There is a convenience wrapper _MM_GET_FLUSH_ZERO_MODE().
-///    </li>
-///    <li>
-///      For checking denormals-are-zero mode: _MM_DENORMALS_ZERO_ON,
-///      _MM_DENORMALS_ZERO_OFF. There is a convenience wrapper
-///      _MM_GET_DENORMALS_ZERO_MODE().
-///    </li>
-///    </ul>
-///
-///    For example, the following expression checks if an overflow exception has
-///    occurred:
-///    \code
-///      ( _mm_getcsr() & _MM_EXCEPT_OVERFLOW )
-///    \endcode
-///
-///    The following expression gets the current rounding mode:
-///    \code
-///      _MM_GET_ROUNDING_MODE()
-///    \endcode
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VSTMXCSR / STMXCSR </c> instruction.
-///
-/// \returns A 32-bit unsigned integer containing the contents of the MXCSR
-///    register.
-unsigned int _mm_getcsr(void);
-
-/// Sets the MXCSR register with the 32-bit unsigned integer value.
-///
-///    There are several groups of macros associated with this intrinsic,
-///    including:
-///    <ul>
-///    <li>
-///      For setting exception states: _MM_EXCEPT_INVALID, _MM_EXCEPT_DIV_ZERO,
-///      _MM_EXCEPT_DENORM, _MM_EXCEPT_OVERFLOW, _MM_EXCEPT_UNDERFLOW,
-///      _MM_EXCEPT_INEXACT. There is a convenience wrapper
-///      _MM_SET_EXCEPTION_STATE(x) where x is one of these macros.
-///    </li>
-///    <li>
-///      For setting exception masks: _MM_MASK_UNDERFLOW, _MM_MASK_OVERFLOW,
-///      _MM_MASK_INVALID, _MM_MASK_DENORM, _MM_MASK_DIV_ZERO, _MM_MASK_INEXACT.
-///      There is a convenience wrapper _MM_SET_EXCEPTION_MASK(x) where x is one
-///      of these macros.
-///    </li>
-///    <li>
-///      For setting rounding modes: _MM_ROUND_NEAREST, _MM_ROUND_DOWN,
-///      _MM_ROUND_UP, _MM_ROUND_TOWARD_ZERO. There is a convenience wrapper
-///      _MM_SET_ROUNDING_MODE(x) where x is one of these macros.
-///    </li>
-///    <li>
-///      For setting flush-to-zero mode: _MM_FLUSH_ZERO_ON, _MM_FLUSH_ZERO_OFF.
-///      There is a convenience wrapper _MM_SET_FLUSH_ZERO_MODE(x) where x is
-///      one of these macros.
-///    </li>
-///    <li>
-///      For setting denormals-are-zero mode: _MM_DENORMALS_ZERO_ON,
-///      _MM_DENORMALS_ZERO_OFF. There is a convenience wrapper
-///      _MM_SET_DENORMALS_ZERO_MODE(x) where x is one of these macros.
-///    </li>
-///    </ul>
-///
-///    For example, the following expression causes subsequent floating-point
-///    operations to round up:
-///      _mm_setcsr(_mm_getcsr() | _MM_ROUND_UP)
-///
-///    The following example sets the DAZ and FTZ flags:
-///    \code
-///    void setFlags() {
-///      _MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON);
-///      _MM_SET_DENORMALS_ZERO_MODE(_MM_DENORMALS_ZERO_ON);
-///    }
-///    \endcode
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VLDMXCSR / LDMXCSR </c> instruction.
-///
-/// \param __i
-///    A 32-bit unsigned integer value to be written to the MXCSR register.
-void _mm_setcsr(unsigned int __i);
-
-#if defined(__cplusplus)
-} // extern "C"
-#endif
-
-/// Selects 4 float values from the 128-bit operands of [4 x float], as
-///    specified by the immediate value operand.
-///
-/// \headerfile <x86intrin.h>
-///
-/// \code
-/// __m128 _mm_shuffle_ps(__m128 a, __m128 b, const int mask);
-/// \endcode
-///
-/// This intrinsic corresponds to the <c> VSHUFPS / SHUFPS </c> instruction.
-///
-/// \param a
-///    A 128-bit vector of [4 x float].
-/// \param b
-///    A 128-bit vector of [4 x float].
-/// \param mask
-///    An immediate value containing an 8-bit value specifying which elements to
-///    copy from \a a and \a b. \n
-///    Bits [3:0] specify the values copied from operand \a a. \n
-///    Bits [7:4] specify the values copied from operand \a b. \n
-///    The destinations within the 128-bit destination are assigned values as
-///    follows: \n
-///    Bits [1:0] are used to assign values to bits [31:0] in the
-///    destination. \n
-///    Bits [3:2] are used to assign values to bits [63:32] in the
-///    destination. \n
-///    Bits [5:4] are used to assign values to bits [95:64] in the
-///    destination. \n
-///    Bits [7:6] are used to assign values to bits [127:96] in the
-///    destination. \n
-///    Bit value assignments: \n
-///    00: Bits [31:0] copied from the specified operand. \n
-///    01: Bits [63:32] copied from the specified operand. \n
-///    10: Bits [95:64] copied from the specified operand. \n
-///    11: Bits [127:96] copied from the specified operand.
-/// \returns A 128-bit vector of [4 x float] containing the shuffled values.
-#define _mm_shuffle_ps(a, b, mask) \
-  (__m128)__builtin_ia32_shufps((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), \
-                                (int)(mask))
-
-/// Unpacks the high-order (index 2,3) values from two 128-bit vectors of
-///    [4 x float] and interleaves them into a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKHPS / UNPCKHPS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. \n
-///    Bits [95:64] are written to bits [31:0] of the destination. \n
-///    Bits [127:96] are written to bits [95:64] of the destination.
-/// \param __b
-///    A 128-bit vector of [4 x float].
-///    Bits [95:64] are written to bits [63:32] of the destination. \n
-///    Bits [127:96] are written to bits [127:96] of the destination.
-/// \returns A 128-bit vector of [4 x float] containing the interleaved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_unpackhi_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 2, 6, 3, 7);
-}
-
-/// Unpacks the low-order (index 0,1) values from two 128-bit vectors of
-///    [4 x float] and interleaves them into a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKLPS / UNPCKLPS </c> instruction.
-///
-/// \param __a
-///    A 128-bit vector of [4 x float]. \n
-///    Bits [31:0] are written to bits [31:0] of the destination.  \n
-///    Bits [63:32] are written to bits [95:64] of the destination.
-/// \param __b
-///    A 128-bit vector of [4 x float]. \n
-///    Bits [31:0] are written to bits [63:32] of the destination. \n
-///    Bits [63:32] are written to bits [127:96] of the destination.
-/// \returns A 128-bit vector of [4 x float] containing the interleaved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_unpacklo_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 4, 1, 5);
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
-///    32 bits are set to the lower 32 bits of the second parameter. The upper
-///    96 bits are set to the upper 96 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VBLENDPS / BLENDPS / MOVSS </c>
-///    instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float]. The upper 96 bits are
-///    written to the upper 96 bits of the result.
-/// \param __b
-///    A 128-bit floating-point vector of [4 x float]. The lower 32 bits are
-///    written to the lower 32 bits of the result.
-/// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_move_ss(__m128 __a, __m128 __b)
-{
-  __a[0] = __b[0];
-  return __a;
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
-///    64 bits are set to the upper 64 bits of the second parameter. The upper
-///    64 bits are set to the upper 64 bits of the first parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKHPD / UNPCKHPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float]. The upper 64 bits are
-///    written to the upper 64 bits of the result.
-/// \param __b
-///    A 128-bit floating-point vector of [4 x float]. The upper 64 bits are
-///    written to the lower 64 bits of the result.
-/// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_movehl_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 6, 7, 2, 3);
-}
-
-/// Constructs a 128-bit floating-point vector of [4 x float]. The lower
-///    64 bits are set to the lower 64 bits of the first parameter. The upper
-///    64 bits are set to the lower 64 bits of the second parameter.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VUNPCKLPD / UNPCKLPD </c> instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float]. The lower 64 bits are
-///    written to the lower 64 bits of the result.
-/// \param __b
-///    A 128-bit floating-point vector of [4 x float]. The lower 64 bits are
-///    written to the upper 64 bits of the result.
-/// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_movelh_ps(__m128 __a, __m128 __b)
-{
-  return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 1, 4, 5);
-}
-
-/// Converts a 64-bit vector of [4 x i16] into a 128-bit vector of [4 x
-///    float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of [4 x i16]. The elements of the destination are copied
-///    from the corresponding elements in this operand.
-/// \returns A 128-bit vector of [4 x float] containing the copied and converted
-///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpi16_ps(__m64 __a)
-{
-  __m64 __b, __c;
-  __m128 __r;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_cmpgt_pi16(__b, __a);
-  __c = _mm_unpackhi_pi16(__a, __b);
-  __r = _mm_setzero_ps();
-  __r = _mm_cvtpi32_ps(__r, __c);
-  __r = _mm_movelh_ps(__r, __r);
-  __c = _mm_unpacklo_pi16(__a, __b);
-  __r = _mm_cvtpi32_ps(__r, __c);
-
-  return __r;
-}
-
-/// Converts a 64-bit vector of 16-bit unsigned integer values into a
-///    128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of 16-bit unsigned integer values. The elements of the
-///    destination are copied from the corresponding elements in this operand.
-/// \returns A 128-bit vector of [4 x float] containing the copied and converted
-///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpu16_ps(__m64 __a)
-{
-  __m64 __b, __c;
-  __m128 __r;
-
-  __b = _mm_setzero_si64();
-  __c = _mm_unpackhi_pi16(__a, __b);
-  __r = _mm_setzero_ps();
-  __r = _mm_cvtpi32_ps(__r, __c);
-  __r = _mm_movelh_ps(__r, __r);
-  __c = _mm_unpacklo_pi16(__a, __b);
-  __r = _mm_cvtpi32_ps(__r, __c);
-
-  return __r;
-}
-
-/// Converts the lower four 8-bit values from a 64-bit vector of [8 x i8]
-///    into a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of [8 x i8]. The elements of the destination are copied
-///    from the corresponding lower 4 elements in this operand.
-/// \returns A 128-bit vector of [4 x float] containing the copied and converted
-///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpi8_ps(__m64 __a)
-{
-  __m64 __b;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_cmpgt_pi8(__b, __a);
-  __b = _mm_unpacklo_pi8(__a, __b);
-
-  return _mm_cvtpi16_ps(__b);
-}
-
-/// Converts the lower four unsigned 8-bit integer values from a 64-bit
-///    vector of [8 x u8] into a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of unsigned 8-bit integer values. The elements of the
-///    destination are copied from the corresponding lower 4 elements in this
-///    operand.
-/// \returns A 128-bit vector of [4 x float] containing the copied and converted
-///    values from the source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpu8_ps(__m64 __a)
-{
-  __m64 __b;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_unpacklo_pi8(__a, __b);
-
-  return _mm_cvtpi16_ps(__b);
-}
-
-/// Converts the two 32-bit signed integer values from each 64-bit vector
-///    operand of [2 x i32] into a 128-bit vector of [4 x float].
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPI2PS + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 64-bit vector of [2 x i32]. The lower elements of the destination are
-///    copied from the elements in this operand.
-/// \param __b
-///    A 64-bit vector of [2 x i32]. The upper elements of the destination are
-///    copied from the elements in this operand.
-/// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
-///    copied and converted values from the first operand. The upper 64 bits
-///    contain the copied and converted values from the second operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
-{
-  __m128 __c;
-
-  __c = _mm_setzero_ps();
-  __c = _mm_cvtpi32_ps(__c, __b);
-  __c = _mm_movelh_ps(__c, __c);
-
-  return _mm_cvtpi32_ps(__c, __a);
-}
-
-/// Converts each single-precision floating-point element of a 128-bit
-///    floating-point vector of [4 x float] into a 16-bit signed integer, and
-///    packs the results into a 64-bit integer vector of [4 x i16].
-///
-///    If the floating-point element is NaN or infinity, or if the
-///    floating-point element is greater than 0x7FFFFFFF or less than -0x8000,
-///    it is converted to 0x8000. Otherwise if the floating-point element is
-///    greater than 0x7FFF, it is converted to 0x7FFF.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPS2PI + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float].
-/// \returns A 64-bit integer vector of [4 x i16] containing the converted
-///    values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtps_pi16(__m128 __a)
-{
-  __m64 __b, __c;
-
-  __b = _mm_cvtps_pi32(__a);
-  __a = _mm_movehl_ps(__a, __a);
-  __c = _mm_cvtps_pi32(__a);
-
-  return _mm_packs_pi32(__b, __c);
-}
-
-/// Converts each single-precision floating-point element of a 128-bit
-///    floating-point vector of [4 x float] into an 8-bit signed integer, and
-///    packs the results into the lower 32 bits of a 64-bit integer vector of
-///    [8 x i8]. The upper 32 bits of the vector are set to 0.
-///
-///    If the floating-point element is NaN or infinity, or if the
-///    floating-point element is greater than 0x7FFFFFFF or less than -0x80, it
-///    is converted to 0x80. Otherwise if the floating-point element is greater
-///    than 0x7F, it is converted to 0x7F.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> CVTPS2PI + COMPOSITE </c> instruction.
-///
-/// \param __a
-///    128-bit floating-point vector of [4 x float].
-/// \returns A 64-bit integer vector of [8 x i8]. The lower 32 bits contain the
-///    converted values and the uppper 32 bits are set to zero.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
-_mm_cvtps_pi8(__m128 __a)
-{
-  __m64 __b, __c;
-
-  __b = _mm_cvtps_pi16(__a);
-  __c = _mm_setzero_si64();
-
-  return _mm_packs_pi16(__b, __c);
-}
-
-/// Extracts the sign bits from each single-precision floating-point
-///    element of a 128-bit floating-point vector of [4 x float] and returns the
-///    sign bits in bits [0:3] of the result. Bits [31:4] of the result are set
-///    to zero.
-///
-/// \headerfile <x86intrin.h>
-///
-/// This intrinsic corresponds to the <c> VMOVMSKPS / MOVMSKPS </c> instruction.
-///
-/// \param __a
-///    A 128-bit floating-point vector of [4 x float].
-/// \returns A 32-bit integer value. Bits [3:0] contain the sign bits from each
-///    single-precision floating-point element of the parameter. Bits [31:4] are
-///    set to zero.
-static __inline__ int __DEFAULT_FN_ATTRS
-_mm_movemask_ps(__m128 __a)
-{
-  return __builtin_ia32_movmskps((__v4sf)__a);
-}
-
-
-#define _MM_ALIGN16 __attribute__((aligned(16)))
-
-#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))
-
-#define _MM_EXCEPT_INVALID    (0x0001)
-#define _MM_EXCEPT_DENORM     (0x0002)
-#define _MM_EXCEPT_DIV_ZERO   (0x0004)
-#define _MM_EXCEPT_OVERFLOW   (0x0008)
-#define _MM_EXCEPT_UNDERFLOW  (0x0010)
-#define _MM_EXCEPT_INEXACT    (0x0020)
-#define _MM_EXCEPT_MASK       (0x003f)
-
-#define _MM_MASK_INVALID      (0x0080)
-#define _MM_MASK_DENORM       (0x0100)
-#define _MM_MASK_DIV_ZERO     (0x0200)
-#define _MM_MASK_OVERFLOW     (0x0400)
-#define _MM_MASK_UNDERFLOW    (0x0800)
-#define _MM_MASK_INEXACT      (0x1000)
-#define _MM_MASK_MASK         (0x1f80)
-
-#define _MM_ROUND_NEAREST     (0x0000)
-#define _MM_ROUND_DOWN        (0x2000)
-#define _MM_ROUND_UP          (0x4000)
-#define _MM_ROUND_TOWARD_ZERO (0x6000)
-#define _MM_ROUND_MASK        (0x6000)
-
-#define _MM_FLUSH_ZERO_MASK   (0x8000)
-#define _MM_FLUSH_ZERO_ON     (0x8000)
-#define _MM_FLUSH_ZERO_OFF    (0x0000)
-
-#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
-#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
-#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
-#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)
-
-#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
-#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
-#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
-#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))
-
-#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
-do { \
-  __m128 tmp3, tmp2, tmp1, tmp0; \
-  tmp0 = _mm_unpacklo_ps((row0), (row1)); \
-  tmp2 = _mm_unpacklo_ps((row2), (row3)); \
-  tmp1 = _mm_unpackhi_ps((row0), (row1)); \
-  tmp3 = _mm_unpackhi_ps((row2), (row3)); \
-  (row0) = _mm_movelh_ps(tmp0, tmp2); \
-  (row1) = _mm_movehl_ps(tmp2, tmp0); \
-  (row2) = _mm_movelh_ps(tmp1, tmp3); \
-  (row3) = _mm_movehl_ps(tmp3, tmp1); \
-} while (0)
-
-/* Aliases for compatibility. */
-#define _m_pextrw _mm_extract_pi16
-#define _m_pinsrw _mm_insert_pi16
-#define _m_pmaxsw _mm_max_pi16
-#define _m_pmaxub _mm_max_pu8
-#define _m_pminsw _mm_min_pi16
-#define _m_pminub _mm_min_pu8
-#define _m_pmovmskb _mm_movemask_pi8
-#define _m_pmulhuw _mm_mulhi_pu16
-#define _m_pshufw _mm_shuffle_pi16
-#define _m_maskmovq _mm_maskmove_si64
-#define _m_pavgb _mm_avg_pu8
-#define _m_pavgw _mm_avg_pu16
-#define _m_psadbw _mm_sad_pu8
-#define _m_ _mm_
-#define _m_ _mm_
-
-#undef __DEFAULT_FN_ATTRS
-#undef __DEFAULT_FN_ATTRS_MMX
-
-/* Ugly hack for backwards-compatibility (compatible with gcc) */
-#if defined(__SSE2__) && !__building_module(_Builtin_intrinsics)
-#include <emmintrin.h>
-#endif
-
-#endif /* __XMMINTRIN_H */
diff --git a/linux-x86/lib64/libLLVM-10svn.so b/linux-x86/lib64/libLLVM-10svn.so
new file mode 100644
index 0000000..f1d842b
--- /dev/null
+++ b/linux-x86/lib64/libLLVM-10svn.so
Binary files differ
diff --git a/linux-x86/lib64/libLLVM-9svn.so b/linux-x86/lib64/libLLVM-9svn.so
deleted file mode 100644
index 6588c6a..0000000
--- a/linux-x86/lib64/libLLVM-9svn.so
+++ /dev/null
Binary files differ
diff --git a/linux-x86/lib64/libbase.so b/linux-x86/lib64/libbase.so
index 2a05f72..ecac42c 100755
--- a/linux-x86/lib64/libbase.so
+++ b/linux-x86/lib64/libbase.so
Binary files differ
diff --git a/linux-x86/lib64/libc++.so b/linux-x86/lib64/libc++.so
index bad356d..15bd1b4 100755
--- a/linux-x86/lib64/libc++.so
+++ b/linux-x86/lib64/libc++.so
Binary files differ
diff --git a/linux-x86/lib64/libc++.so.1 b/linux-x86/lib64/libc++.so.1
index 8795469..0f28ca5 100644
--- a/linux-x86/lib64/libc++.so.1
+++ b/linux-x86/lib64/libc++.so.1
Binary files differ
diff --git a/linux-x86/lib64/libclang_cxx.so.10svn b/linux-x86/lib64/libclang_cxx.so.10svn
new file mode 100644
index 0000000..b071f5b
--- /dev/null
+++ b/linux-x86/lib64/libclang_cxx.so.10svn
Binary files differ
diff --git a/linux-x86/lib64/libclang_cxx.so.9svn b/linux-x86/lib64/libclang_cxx.so.9svn
deleted file mode 100644
index 3a77c72..0000000
--- a/linux-x86/lib64/libclang_cxx.so.9svn
+++ /dev/null
Binary files differ
diff --git a/linux-x86/lib64/liblog.so b/linux-x86/lib64/liblog.so
index b19a9fb..3837cc2 100755
--- a/linux-x86/lib64/liblog.so
+++ b/linux-x86/lib64/liblog.so
Binary files differ
diff --git a/linux-x86/lib64/libprotobuf-cpp-full.so b/linux-x86/lib64/libprotobuf-cpp-full.so
index dc2f493..88d5e40 100755
--- a/linux-x86/lib64/libprotobuf-cpp-full.so
+++ b/linux-x86/lib64/libprotobuf-cpp-full.so
Binary files differ
diff --git a/linux-x86/lib64/libz-host.so b/linux-x86/lib64/libz-host.so
index 30aea85..8d8d0af 100755
--- a/linux-x86/lib64/libz-host.so
+++ b/linux-x86/lib64/libz-host.so
Binary files differ
diff --git a/linux-x86/lib64/libziparchive.so b/linux-x86/lib64/libziparchive.so
index 93b392a..bfdcc32 100755
--- a/linux-x86/lib64/libziparchive.so
+++ b/linux-x86/lib64/libziparchive.so
Binary files differ
diff --git a/manifest.xml b/manifest.xml
index dbecec3..d34ec2f 100644
--- a/manifest.xml
+++ b/manifest.xml
@@ -5,103 +5,109 @@
 
   <default remote="aosp" revision="master" />
 
-  <project name="platform/development" path="development" revision="6e2e77336e316737ddc7fe52d77945db6e0205be" />
+  <project name="platform/development" path="development" revision="eaee8a52291904dc83b4ea3b72181fa373ffd92e" />
 
-  <project clone-depth="1" name="platform/prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.17-4.8" path="prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.17-4.8" revision="639243dd5cb2c5ea1313f033c75c8c533813a484" />
+  <project clone-depth="1" name="platform/prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.17-4.8" path="prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.17-4.8" revision="faf5c0c0ba267a56ee16a88bc8983ec9668229e6" />
 
-  <project name="platform/build/soong" path="build/soong" revision="af60d490ff6fdd716b343e0b29ce5b2bb95e7829">
+  <project name="platform/external/go-creachadair-stringset" path="external/go-creachadair-stringset" revision="580fe893c0cd9bc1cb41a735b6ff456901f00e83" />
+
+  <project name="platform/build/soong" path="build/soong" revision="e35ff4aeb32f49a7f20f2367aee53c575bf45000">
     <linkfile dest="Android.bp" src="root.bp" />
 
     <linkfile dest="bootstrap.bash" src="bootstrap.bash" />
 </project>
 
-  <project name="platform/external/ninja" path="external/ninja" revision="059335d867310d6e8bf9fa44de69755575e763d2" />
+  <project name="platform/external/ninja" path="external/ninja" revision="98d29d1997d682d68525df68c7e78b25fee828fc" />
+
+  <project name="platform/build/blueprint" path="build/blueprint" revision="86286b9ac935c3d81d81044cc10d0d436ad09029" />
 
   <project name="platform/external/rapidjson" path="external/rapidjson" revision="9fa2a3d9e356a1f42a6184dcf1e0508ddfa9dbfb" />
 
-  <project name="platform/external/jsoncpp" path="external/jsoncpp" revision="7613e8a4989f43b319a0504444e698f5072eacea" />
+  <project name="platform/external/jsoncpp" path="external/jsoncpp" revision="34ba61cd2766f63917262655c46f4075a3c14712" />
 
-  <project name="platform/external/googletest" path="external/googletest" revision="734b56dc80a5979e921620205b0758844a131d60" />
+  <project name="platform/external/googletest" path="external/googletest" revision="a09ca9ae1c2ef1b305235d905bcd8444f9a8ce44" />
 
   <project name="platform/external/regex-re2" path="external/regex-re2" revision="84e28962b2c2f357b5daccb460501b169193fafe" />
 
-  <project clone-depth="1" name="platform/prebuilts/build-tools" path="prebuilts/build-tools" revision="754018d40bd84ab3ee838d8255652e449c5a7873" />
+  <project clone-depth="1" name="platform/prebuilts/build-tools" path="prebuilts/build-tools" revision="070ee5bbc8c951b524b98040013400a0148de559" />
 
-  <project name="platform/bionic" path="bionic" revision="5d9ba9ea702a4d20b99639996b8bd27f121d5bd7" />
+  <project name="platform/bionic" path="bionic" revision="d2b8149ddad8b21ce535fb02ff752c85c7aa9d43" />
 
-  <project clone-depth="1" groups="linux" name="platform/prebuilts/go/linux-x86" path="prebuilts/go/linux-x86" revision="de981b817157f5ca963456df355b40912132a858" />
+  <project clone-depth="1" groups="linux" name="platform/prebuilts/go/linux-x86" path="prebuilts/go/linux-x86" revision="537e2072e1affb379b4da1ecfbcb41a270544d1a" />
 
-  <project name="platform/external/zlib" path="external/zlib" revision="046b1c12dbfb697696d682d4207994f52300c309" />
+  <project name="platform/external/zlib" path="external/zlib" revision="73131478df9bdc6f99c96f2a33c773eff4b570be" />
+
+  <project name="platform/external/golang-x-tools" path="external/golang-x-tools" revision="bcc6484babb4e999e4bdb6a982bfa7e92fc640d2" />
 
   <project name="platform/external/zopfli" path="external/zopfli" revision="9bdeb8615dc9a4eda08f72c1f81479075a4cb425" />
 
   <project clone-depth="1" name="platform/prebuilts/jdk/jdk11" path="prebuilts/jdk/jdk11" revision="38b397ddfe0c3339623cec843c810bd938de9a89" />
 
-  <project name="platform/external/kythe" path="external/kythe" revision="b2e7760579730215040299db5dd17bd556c3cb26" />
+  <project name="platform/external/kythe" path="external/kythe" revision="d324727d452449bcc3aede7d23ecba0c5589d479" />
 
-  <project name="platform/external/libcxx" path="external/libcxx" revision="14a8db71eb959fd5c2acf49ae716169005af4643" />
+  <project name="platform/external/libcxx" path="external/libcxx" revision="fb268a4f43cfa8edc0bde47c1370405d78605b75" />
 
-  <project groups="pdk" name="platform/external/boringssl" path="external/boringssl" revision="0f2e90fc6fe301beefd2a8c27a5720d84a4e3117" />
+  <project groups="pdk" name="platform/external/boringssl" path="external/boringssl" revision="cdee22e79e64eeebdf467a90c3d06639adee0a45" />
 
   <project clone-depth="1" groups="linux" name="platform/prebuilts/ninja/linux-x86" path="prebuilts/ninja/linux-x86" revision="6369b19fc3fbe765636af75d394627e2b92599ed" />
 
-  <project name="platform/external/clang" path="external/clang" revision="8343f599c9e1dda8e6c1bb779af323a673a052f9" />
+  <project name="platform/external/clang" path="external/clang" revision="4c9ce042c54a4bccb4f0fc36e605e3df5be5a296" />
 
-  <project name="platform/build/kati" path="build/kati" revision="f286cf8c3ec14128b8d6396870580fcc3bbdd96b" />
+  <project name="platform/external/libcxxabi" path="external/libcxxabi" revision="8ba49a025b6ee7d2ea78e68b0a3b29181e0b2149" />
 
   <project clone-depth="1" name="platform/prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.15-4.8" path="prebuilts/gcc/linux-x86/host/x86_64-linux-glibc2.15-4.8" revision="d9aafaade740ca38612c742f6d87debf362132ea" />
 
-  <project name="platform/external/libcxxabi" path="external/libcxxabi" revision="8920c61b6bfd666e1c45c1f2b8a2ebc3400db648" />
+  <project name="platform/build/kati" path="build/kati" revision="32c653b31bd1a6ce54ebbb1b73ce03bde6c92246" />
 
   <project clone-depth="1" groups="darwin" name="platform/prebuilts/gcc/darwin-x86/host/headers" path="prebuilts/gcc/darwin-x86/host/headers" revision="4ac4f7cc41cf3c9e36fc3d6cf37fd1cfa9587a68" />
 
   <project name="platform/external/libunwind" path="external/libunwind" revision="2f060ec263cd6fa3ea56140c84439d08e360621d" />
 
-  <project name="platform/system/core" path="system/core" revision="4b3e5e1de82efda8ba2ac1aba311dac9befed6a4" />
+  <project name="platform/system/core" path="system/core" revision="3bf88581050eed7151c63bfe871f3a7708e31aaa" />
 
-  <project clone-depth="1" groups="linux" name="platform/prebuilts/clang/host/linux-x86" path="prebuilts/clang/host/linux-x86" revision="83d401b133d301508c2a8e0f9afccc8dbcea5f08" />
+  <project clone-depth="1" groups="linux" name="platform/prebuilts/clang/host/linux-x86" path="prebuilts/clang/host/linux-x86" revision="5ce3025475e53d62be50f2b1b67cd0385e9328a6" />
 
-  <project clone-depth="3" name="platform/prebuilts/clang-tools" path="prebuilts/clang-tools" revision="367167e08119c148815dd88446a611fbc3899e04" />
+  <project clone-depth="3" name="platform/prebuilts/clang-tools" path="prebuilts/clang-tools" revision="ef9135150a3cfddbaad88bcd30ade0cecb0ac41c" />
 
-  <project name="platform/external/golang-protobuf" path="external/golang-protobuf" revision="6665b9a881c57e344ff785850e528f833982cea1" />
+  <project name="platform/external/golang-protobuf" path="external/golang-protobuf" revision="fcdf38bba2f334cda94ca7a01fef5b3a55307deb" />
 
   <project clone-depth="1" name="platform/prebuilts/jdk/jdk9" path="prebuilts/jdk/jdk9" revision="3447ef32fa4961229942fa345c8707c7f9e00246" />
 
-  <project clone-depth="1" groups="darwin" name="platform/prebuilts/gcc/darwin-x86/host/i686-apple-darwin-4.2.1" path="prebuilts/gcc/darwin-x86/host/i686-apple-darwin-4.2.1" revision="ec5aa66aaa4964c27564d0ec84dc1f18a2d72b7e" />
+  <project clone-depth="1" groups="darwin" name="platform/prebuilts/gcc/darwin-x86/host/i686-apple-darwin-4.2.1" path="prebuilts/gcc/darwin-x86/host/i686-apple-darwin-4.2.1" revision="f596bca2999d9d0e61109659f66cffe1e82aa32e" />
 
-  <project name="platform/external/python/cpython3" path="external/python/cpython3" revision="d3946a97a47cb35880d270d983721883a21b26c1" />
+  <project name="platform/external/python/cpython3" path="external/python/cpython3" revision="5b9fd40b7072410f7bfc60da9c0dba7b9909b192" />
 
-  <project name="platform/external/compiler-rt" path="external/compiler-rt" revision="36f0dec43fd96956ad862f27f562b1b035b138a7" />
+  <project name="platform/external/compiler-rt" path="external/compiler-rt" revision="c72b929327e09ae6c482ff6d091f191c64044c43" />
 
-  <project name="platform/build/blueprint" path="build/blueprint" revision="247a72401f5115b16e4165205cac25edafd757cb" />
+  <project name="platform/external/golang-x-sync" path="external/golang-x-sync" revision="bf7c9e83a86ea08d93262ec468a53e866728e36a" />
 
   <project name="platform/external/libunwind_llvm" path="external/libunwind_llvm" revision="249d6896be0c5179c6d6b4fd1e3d8f092a01582d" />
 
-  <project name="platform/external/fmtlib" path="external/fmtlib" revision="0a50c7bef3c4fdecf1e3798f0031988be4b69688" />
+  <project name="platform/external/fmtlib" path="external/fmtlib" revision="289bd9700073441c2086388c1dc85aa3daea4e08" />
 
-  <project name="platform/external/llvm" path="external/llvm" revision="9b83b879e21e9c4f16b12ac30829462880142774" />
+  <project name="platform/external/llvm" path="external/llvm" revision="428628891fde8aef207fb9302ca8917ff19dd025" />
 
-  <project groups="pdk" name="platform/external/gflags" path="external/gflags" revision="0ff6b6b12c67175a1573050e2ded5f7951b989ea" />
+  <project groups="pdk" name="platform/external/gflags" path="external/gflags" revision="ab750f5fe55da2b363ab745908fe2673967610e9" />
 
-  <project name="platform/build" path="build/make" revision="db49de8560414ae221b61bafd0c244729c5610cb">
+  <project name="platform/build" path="build/make" revision="eb28123fef0f9435ac8b5eea09e26af007f9393c">
     <linkfile dest="build/tools" src="tools" />
 </project>
 
-  <project clone-depth="1" name="platform/prebuilts/misc" path="prebuilts/misc" revision="f094ca06f4378faaa627999efe12913dc2f37b89" />
+  <project clone-depth="1" name="platform/prebuilts/misc" path="prebuilts/misc" revision="046af3688507e097f9ed247fb07d261338d407b7" />
 
-  <project clone-depth="1" groups="darwin" name="platform/prebuilts/clang/host/darwin-x86" path="prebuilts/clang/host/darwin-x86" revision="7bf3c2e36f8d52e6f3914fc5a29ed937657d6e0a" />
+  <project clone-depth="1" groups="darwin" name="platform/prebuilts/clang/host/darwin-x86" path="prebuilts/clang/host/darwin-x86" revision="a19265e4bc3e6695fc406dda4b30488032d27329" />
 
-  <project name="platform/external/protobuf" path="external/protobuf" revision="b6ad51875e54200ae19128e82bd45545b9e71ed6" />
+  <project name="platform/external/protobuf" path="external/protobuf" revision="e6ebc68a04883e79701373894a8eceb142b9c258" />
 
   <project clone-depth="1" name="platform/prebuilts/jdk/jdk8" path="prebuilts/jdk/jdk8" revision="74e4f1844dfa9b8df9e0fe2ff34a2ecc24d52b07" />
 
   <project clone-depth="1" groups="darwin" name="platform/prebuilts/ninja/darwin-x86" path="prebuilts/ninja/darwin-x86" revision="00f798346dedb4a7a6a6dcc9ad32ff09d66ee0db" />
 
-  <project name="platform/external/python/cpython2" path="external/python/cpython2" revision="1c13c899cb54e69f9c795640f981becc7a507298" />
+  <project name="platform/external/python/cpython2" path="external/python/cpython2" revision="8c8f0bb4662076c4acab2123bd8de180f9f8ecaf" />
 
-  <project name="platform/dalvik" path="dalvik" revision="95783d06ddc5390ad88b9870171238dde385d9aa" />
+  <project name="platform/dalvik" path="dalvik" revision="54cc4ca1beb7eba4509005553e89ad1ea4528812" />
 
-  <project name="platform/external/abseil-cpp" path="external/abseil-cpp" revision="8192e47b4052c5103408d8a93c0e8a5255a69977" />
+  <project name="platform/external/abseil-cpp" path="external/abseil-cpp" revision="c46f2ec91c41d06f7731120c061de98fd2b86b95" />
 
-  <project clone-depth="1" groups="darwin" name="platform/prebuilts/go/darwin-x86" path="prebuilts/go/darwin-x86" revision="2c9da2e2d268ffd45952a74c916c7c9dd0bd8275" />
+  <project clone-depth="1" groups="darwin" name="platform/prebuilts/go/darwin-x86" path="prebuilts/go/darwin-x86" revision="1a68b8d2b99293ce55cd444bb764abdd29100463" />
 </manifest>